4.7-alpha2 (#1027)

* feat: stop toolCall and rename some field. (#46) * perf: node delete tip;pay tip * fix: toolCall cannot save child answer * feat: stop tool * fix: team modal * fix feckbackMoal auth bug (#47) * 简单的支持提示词运行tool。优化workflow模板 (#49) * remove templates * fix: request body undefined * feat: prompt tool run * feat: workflow tamplates modal * perf: plugin start * 4.7 (#50) * fix docker-compose download url (#994) original code is a bad url with '404 NOT FOUND' return. fix docker-compose download url, add 'v' before docker-compose version * Update ai_settings.md (#1000) * Update configuration.md * Update configuration.md * Fix history in classifyQuestion and extract modules (#1012) * Fix history in classifyQuestion and extract modules * Add chatValue2RuntimePrompt import and update text formatting * flow controller to packages * fix: rerank select * modal ui * perf: modal code path * point not sufficient * feat: http url support variable * fix http key * perf: prompt * perf: ai setting modal * simple edit ui --------- Co-authored-by: entorick <entorick11@qq.com> Co-authored-by: liujianglc <liujianglc@163.com> Co-authored-by: Fengrui Liu <liufengrui.work@bytedance.com> * fix team share redirect to login (#51) * feat: support openapi import plugins (#48) * feat: support openapi import plugins * feat: import from url * fix: add body params parse * fix build * fix * fix * fix * tool box ui (#52) * fix: training queue * feat: simple edit tool select * perf: simple edit dataset prompt * fix: chatbox tool ux * feat: quote prompt module * perf: plugin tools sign * perf: model avatar * tool selector ui * feat: max histories * perf: http plugin import (#53) * perf: plugin http import * chatBox ui * perf: name * fix: Node template card (#54) * fix: ts * setting modal * package * package * feat: add plugins search (#57) * feat: add plugins search * perf: change http plugin header input * Yjl (#56) * perf: prompt tool call * perf: chat box ux * doc * doc * price tip * perf: tool selector * ui' * fix: vector queue * fix: empty tool and empty response * fix: empty msg * perf: pg index * perf: ui tip * doc * tool tip --------- Co-authored-by: yst <77910600+yu-and-liu@users.noreply.github.com> Co-authored-by: entorick <entorick11@qq.com> Co-authored-by: liujianglc <liujianglc@163.com> Co-authored-by: Fengrui Liu <liufengrui.work@bytedance.com> Co-authored-by: heheer <71265218+newfish-cmyk@users.noreply.github.com>
2026-01-01 03:32:51 +00:00 · 2024-03-21 13:32:31 +08:00 · 2024-03-21 13:32:31 +08:00 · 9d27de154b
parent 6d4b331db9
commit 9d27de154b
322 changed files with 9282 additions and 6498 deletions
--- a/README.md
+++ b/README.md
@ -58,7 +58,7 @@ fastgpt.run 域名会弃用。
   - [x] 源文件引用追踪
   - [x] 模块封装，实现多级复用
   - [x] 混合检索 & 重排
-   - [ ] Tool 模块
+   - [x] Tool 模块
   - [ ] 嵌入 [Laf](https://github.com/labring/laf)，实现在线编写 HTTP 模块
   - [ ] 插件封装功能

@ -114,7 +114,7 @@ fastgpt.run 域名会弃用。
 * [多模型配置](https://doc.fastgpt.in/docs/development/one-api/)
 * [版本更新/升级介绍](https://doc.fastgpt.in/docs/development/upgrading)
 * [OpenAPI API 文档](https://doc.fastgpt.in/docs/development/openapi/)
-* [知识库结构详解](https://doc.fastgpt.in/docs/use-cases/datasetengine/)
+* [知识库结构详解](https://doc.fastgpt.in/docs/course/datasetengine/)

 <a href="#readme">
    <img src="https://img.shields.io/badge/-返回顶部-7d09f1.svg" alt="#" align="right">
@ -122,9 +122,9 @@ fastgpt.run 域名会弃用。

 ## 🏘️ 社区交流群

-添加 wx 小助手加入：
+wx 扫一下加入：

-![](https://otnvvf-imgs.oss.laf.run/wx300.jpg)
+![](https://oss.laf.run/htr4n1-images/fastgpt-qr-code.jpg)

 <a href="#readme">
    <img src="https://img.shields.io/badge/-返回顶部-7d09f1.svg" alt="#" align="right">
--- a/README_en.md
+++ b/README_en.md
@ -116,14 +116,12 @@ Project tech stack: NextJs + TS + ChakraUI + Mongo + Postgres (Vector plugin)
 - [Configuring Multiple Models](https://doc.fastgpt.in/docs/installation/reference/models)
 - [Version Updates & Upgrades](https://doc.fastgpt.in/docs/installation/upgrading)

-<!-- ## :point_right: RoadMap
- [FastGPT RoadMap](https://kjqvjse66l.feishu.cn/docx/RVUxdqE2WolDYyxEKATcM0XXnte) -->

-<!-- ## 🏘️ Community
+## 🏘️ Community

-| Community Group                                   | Assistant                                      |
-| ------------------------------------------------- | ---------------------------------------------- |
-| ![](https://otnvvf-imgs.oss.laf.run/wxqun300.jpg) | ![](https://otnvvf-imgs.oss.laf.run/wx300.jpg) | -->
+| Community Group                                   | 
+| ------------------------------------------------- | 
+| ![](https://oss.laf.run/htr4n1-images/fastgpt-qr-code.jpg) | 

 <a href="#readme">
    <img src="https://img.shields.io/badge/-Back_to_Top-7d09f1.svg" alt="#" align="right">
--- a/docSite/assets/imgs/gapierTool1.png
+++ b/docSite/assets/imgs/gapierTool1.png
--- a/docSite/assets/imgs/gapierTool10.png
+++ b/docSite/assets/imgs/gapierTool10.png
--- a/docSite/assets/imgs/gapierTool11.png
+++ b/docSite/assets/imgs/gapierTool11.png
--- a/docSite/assets/imgs/gapierTool12.png
+++ b/docSite/assets/imgs/gapierTool12.png
--- a/docSite/assets/imgs/gapierTool13.png
+++ b/docSite/assets/imgs/gapierTool13.png
--- a/docSite/assets/imgs/gapierTool14.png
+++ b/docSite/assets/imgs/gapierTool14.png
--- a/docSite/assets/imgs/gapierTool15.png
+++ b/docSite/assets/imgs/gapierTool15.png
--- a/docSite/assets/imgs/gapierTool16.png
+++ b/docSite/assets/imgs/gapierTool16.png
--- a/docSite/assets/imgs/gapierTool17.png
+++ b/docSite/assets/imgs/gapierTool17.png
--- a/docSite/assets/imgs/gapierTool2.png
+++ b/docSite/assets/imgs/gapierTool2.png
--- a/docSite/assets/imgs/gapierTool3.png
+++ b/docSite/assets/imgs/gapierTool3.png
--- a/docSite/assets/imgs/gapierTool4.png
+++ b/docSite/assets/imgs/gapierTool4.png
--- a/docSite/assets/imgs/gapierTool5.png
+++ b/docSite/assets/imgs/gapierTool5.png
--- a/docSite/assets/imgs/gapierTool6.png
+++ b/docSite/assets/imgs/gapierTool6.png
--- a/docSite/assets/imgs/gapierTool7.png
+++ b/docSite/assets/imgs/gapierTool7.png
--- a/docSite/assets/imgs/gapierTool8.png
+++ b/docSite/assets/imgs/gapierTool8.png
--- a/docSite/assets/imgs/gapierTool9.png
+++ b/docSite/assets/imgs/gapierTool9.png
--- a/docSite/assets/imgs/gapierToolResult1.png
+++ b/docSite/assets/imgs/gapierToolResult1.png
--- a/docSite/content/docs/commercial/intro.md
+++ b/docSite/content/docs/commercial/intro.md
@ -19,16 +19,17 @@ FastGPT 商业版是基于 FastGPT 开源版的增强版本，增加了一些独
 | 应用管理与高级编排 | ✅ | ✅ | ✅ |
 | 文档知识库 | ✅ | ✅ | ✅ |
 | 外部使用 | ✅ | ✅ | ✅ |
-| 自定义版权信息 | ❌ | ✅ | ✅ |
+| 自定义版权信息 | ❌ | ✅ | 设计中 |
 | 多租户与支付 | ❌ | ✅ | ✅ |
 | 团队空间 | ❌ | ✅ | ✅ |
-| 外部使用限制 | ❌ | ✅ | ✅ |
+| 应用发布安全配置 | ❌ | ✅ | ✅ |
 | 内容审核 | ❌ | ✅ | ✅ |
 | web站点同步 | ❌ | ✅ | ✅ |
 | 管理后台 | ❌ | ✅ | ✅ |
-| Saas服务商业授权 | ❌ | ✅ | ✅ |
+| 完整商业授权 | ❌ | ✅ | ✅ |
 | 图片知识库 | ❌ | 设计中 | 设计中 |
 | 自动规划召回 | ❌ | 设计中 | 设计中 |
+| 对话日志运营分析 | ❌ | 设计中 | 设计中 |
 {{< /table >}}

 ## 商业版软件价格
--- a/docSite/content/docs/use-cases/ai_settings.md
+++ b/docSite/content/docs/use-cases/ai_settings.md
@ -4,7 +4,7 @@ description: "FastGPT AI 高级配置说明"
 icon: "sign_language"
 draft: false
 toc: true
-weight: 501
+weight: 102
 ---

 在 FastGPT 的 AI 对话模块中，有一个 AI 高级配置，里面包含了 AI 模型的参数配置，本文详细介绍这些配置的含义。
@ -48,7 +48,7 @@ Tips: 可以通过点击上下文按键查看完整的上下文组成，便于

 FastGPT 知识库采用 QA 对(不一定都是问答格式，仅代表两个变量)的格式存储，在转义成字符串时候会根据**引用模板**来进行格式化。知识库包含多个可用变量： q, a, sourceId（数据的ID）, index(第n个数据), source(数据的集合名、文件名)，score(距离得分，0-1) 可以通过 {{q}} {{a}} {{sourceId}} {{index}} {{source}} {{score}} 按需引入。下面一个模板例子：

-可以通过 [知识库结构讲解](/docs/use-cases/datasetEngine/) 了解详细的知识库的结构。
+可以通过 [知识库结构讲解](/docs/course/datasetEngine/) 了解详细的知识库的结构。

 #### 引用模板

--- a/docSite/content/docs/use-cases/datasetEngine.md
+++ b/docSite/content/docs/use-cases/datasetEngine.md
@ -4,7 +4,7 @@ description: "本节会详细介绍 FastGPT 知识库结构设计，理解其 QA
 icon: "dataset"
 draft: false
 toc: true
-weight: 502
+weight: 102
 ---

 ## 理解向量
@ -90,4 +90,4 @@ FastGPT 采用了 `PostgresSQL` 的 `PG Vector` 插件作为向量检索器，

 ## QA的组合与引用提示词构建

-参考[引用模板与引用提示词示例](/docs/use-cases/ai_settings/#示例)
+参考[引用模板与引用提示词示例](/docs/course/ai_settings/#示例)
--- a/docSite/content/docs/use-cases/kb.md
+++ b/docSite/content/docs/use-cases/kb.md
@ -4,7 +4,7 @@ description: " 利用 FastGPT 打造高质量 AI 知识库"
 icon: "school"
 draft: false
 toc: true
-weight: 699
+weight: 106
 ---

 ## 前言
--- a/docSite/content/docs/development/configuration.md
+++ b/docSite/content/docs/development/configuration.md
@ -11,7 +11,7 @@ weight: 708

 **开发环境下**，你需要将示例配置文件 `config.json` 复制成 `config.local.json` 文件才会生效。

-这个配置文件中包含了系统参数和各个模型配置，使用时务必去掉注释。
+这个配置文件中包含了系统参数和各个模型配置，`使用时务必去掉注释!!!!!!!!!!!!!!`

 ## 4.6.8+ 版本新配置文件

@ -28,6 +28,7 @@ llm模型全部合并
    {
      "model": "gpt-3.5-turbo", // 模型名
      "name": "gpt-3.5-turbo", // 别名
+      "avatar": "/imgs/model/openai.svg", // 模型的logo
      "maxContext": 16000, // 最大上下文
      "maxResponse": 4000, // 最大回复
      "quoteMaxToken": 13000, // 最大引用内容
@ -35,7 +36,7 @@ llm模型全部合并
      "charsPointsPrice": 0, 
      "censor": false,
      "vision": false, // 是否支持图片输入
-      "datasetProcess": false, // 是否设置为知识库处理模型（QA），务必保证至少有一个为true，否则知识库会报错
+      "datasetProcess": true, // 是否设置为知识库处理模型（QA），务必保证至少有一个为true，否则知识库会报错
      "usedInClassify": true, // 是否用于问题分类（务必保证至少有一个为true）
      "usedInExtractFields": true, // 是否用于内容提取（务必保证至少有一个为true）
      "usedInToolCall": true, // 是否用于工具调用（务必保证至少有一个为true）
@ -47,31 +48,10 @@ llm模型全部合并
      "defaultSystemChatPrompt": "", // 对话默认携带的系统提示词
      "defaultConfig":{}  // LLM默认配置，可以针对不同模型设置特殊值（比如 GLM4 的 top_p
    },
-    {
-      "model": "gpt-3.5-turbo-16k",
-      "name": "gpt-3.5-turbo-16k",
-      "maxContext": 16000,
-      "maxResponse": 16000,
-      "quoteMaxToken": 13000,
-      "maxTemperature": 1.2,
-      "charsPointsPrice": 0,
-      "censor": false,
-      "vision": false,
-      "datasetProcess": true,
-      "usedInClassify": true,
-      "usedInExtractFields": true,
-      "usedInToolCall": true,
-      "usedInQueryExtension": true,
-      "toolChoice": true,
-      "functionCall": false,
-      "customCQPrompt": "",
-      "customExtractPrompt": "",
-      "defaultSystemChatPrompt": "",
-      "defaultConfig":{} 
-    },
    {
      "model": "gpt-4-0125-preview",
      "name": "gpt-4-turbo",
+      "avatar": "/imgs/model/openai.svg",
      "maxContext": 125000,
      "maxResponse": 4000,
      "quoteMaxToken": 100000,
@ -94,6 +74,7 @@ llm模型全部合并
    {
      "model": "gpt-4-vision-preview",
      "name": "gpt-4-vision",
+      "avatar": "/imgs/model/openai.svg",
      "maxContext": 128000,
      "maxResponse": 4000,
      "quoteMaxToken": 100000,
@ -118,6 +99,7 @@ llm模型全部合并
    {
      "model": "text-embedding-ada-002",
      "name": "Embedding-2",
+      "avatar": "/imgs/model/openai.svg",
      "charsPointsPrice": 0,
      "defaultToken": 700,
      "maxToken": 3000,
@ -149,6 +131,20 @@ llm模型全部合并
 }
 ```

+## 关于模型 logo
+
+统一放置在项目的`public/imgs/model/xxx`目录中，目前内置了以下几种，如果有需要，可以PR增加。
+
+- /imgs/model/baichuan.svg - 百川
+- /imgs/model/chatglm.svg - 智谱
+- /imgs/model/calude.svg - calude
+- /imgs/model/ernie.svg - 文心一言
+- /imgs/model/moonshot.svg - 月之暗面
+- /imgs/model/openai.svg - OpenAI GPT
+- /imgs/model/qwen.svg - 通义千问
+- /imgs/model/yi.svg - 零一万物
+- 
+
 ## 特殊模型

 ### ReRank 接入
--- a/docSite/content/docs/development/faq.md
+++ b/docSite/content/docs/development/faq.md
@ -1,13 +1,13 @@
 ---
 weight: 749
-title: "常见开发 & 部署问题"
-description: "FastGPT 常见开发 & 部署问题"
+title: "私有部署常见问题"
+description: "FastGPT 私有部署常见问题"
 icon: upgrade
 draft: false
 images: []
 ---

-## 错误排查方式
+## 一、错误排查方式

 遇到问题先按下面方式排查。

@ -17,7 +17,7 @@ images: []
 4. 无法解决时，可以找找[Issue](https://github.com/labring/FastGPT/issues)，或新提 Issue，私有部署错误，务必提供详细的日志，否则很难排查。


-## 通用问题
+## 二、通用问题

 ### 能否纯本地运行

@ -46,7 +46,7 @@ OneAPI 的 API Key 配置错误，需要修改`OPENAI_API_KEY`环境变量，并
 ### 页面崩溃

 1. 关闭翻译
-2. 检查配置文件是否正常加载，如果没有正常加载会导致缺失系统信息，在某些操作下会导致空指针。（95%情况，可以F12打开控制台，看具体的空指针情况）
+2. 检查配置文件是否正常加载，如果没有正常加载会导致缺失系统信息，在某些操作下会导致空指针。（95%情况是配置文件不对，可以F12打开控制台，看具体的空指针情况）
 3. 某些api不兼容问题（较少）

 ### 开启内容补全后，响应速度变慢
@ -54,7 +54,11 @@ OneAPI 的 API Key 配置错误，需要修改`OPENAI_API_KEY`环境变量，并
 1. 问题补全需要经过一轮AI生成。
 2. 会进行3~5轮的查询，如果数据库性能不足，会有明显影响。

-## 私有部署问题
+### 模型响应为空
+
+1. 检查 key 问题。
+2. 如果是国内模型，可能是命中风控了。
+3. 查看模型请求日志，检查出入参数是否异常。

 ### 知识库索引没有进度

@ -64,11 +68,7 @@ OneAPI 的 API Key 配置错误，需要修改`OPENAI_API_KEY`环境变量，并
 2. 不能对话，也不能索引：API调用失败。可能是没连上OneAPI或OpenAI
 3. 有进度，但是非常慢：api key不行，OpenAI的免费号，一分钟只有3次还是60次。一天上限200次。

-## Docker 部署常见问题
-
-### 首次部署，root用户提示未注册
-
-没有启动 Mongo 副本集模式。
+## 三、Docker 部署常见问题

 ### 如何更新？

@ -133,14 +133,6 @@ mongo连接失败，检查
 2. 环境变量（账号密码，注意host和port）
 3. 副本集启动失败，一直在重启：没挂载mongo key；key没有权限；

-## 本地开发问题
+### 首次部署，root用户提示未注册

-### TypeError: Cannot read properties of null (reading 'useMemo' )
-
-删除所有的`node_modules`，用 Node18 重新 install 试试，可能最新的 Node 有问题。 本地开发流程：
-
-1. 根目录: `pnpm i`
-2. 复制 `config.json` -> `config.local.json`
-3. 复制 `.env.template` -> `.env.local`
-4. `cd projects/app`
-5. `pnpm dev`
+没有启动 Mongo 副本集模式。
--- a/docSite/content/docs/development/intro.md
+++ b/docSite/content/docs/development/intro.md
@ -48,7 +48,7 @@ git clone git@github.com:<github_username>/FastGPT.git

 第一次开发，需要先部署数据库，建议本地开发可以随便找一台 2C2G 的轻量小数据库实践。数据库部署教程：[Docker 快速部署](/docs/development/docker/)。部署完了，可以本地访问其数据库。

-Mongo 数据库需要修改副本集的`host`，从原来的`mongo:27017`修改为`ip:27017`。
+Mongo 数据库需要修改副本集的`host`，从原来的`mongo:27017`修改为`ip:27017`(ip为对应的公网IP)。

 ### 4. 初始配置

@ -113,7 +113,22 @@ docker build -t dockername/fastgpt:tag --build-arg name=app --build-arg proxy=ta

 FastGPT 在`pnpm i`后会执行`postinstall`脚本，用于自动生成`ChakraUI`的`Type`。如果没有权限，可以先执行`chmod -R +x ./scripts/`，再执行`pnpm i`。

-### 加入社区
+### 长时间运行后崩溃
+
+似乎是由于 tiktoken 库的开发环境问题，生产环境中未遇到，暂时可忽略。
+
+### TypeError: Cannot read properties of null (reading 'useMemo' )
+
+删除所有的`node_modules`，用 Node18 重新 install 试试，可能最新的 Node 有问题。 本地开发流程：
+
+1. 根目录: `pnpm i`
+2. 复制 `config.json` -> `config.local.json`
+3. 复制 `.env.template` -> `.env.local`
+4. `cd projects/app`
+5. `pnpm dev`
+
+
+## 加入社区

 遇到困难了吗？有任何问题吗? 加入微信群与开发者和用户保持沟通。

--- a/docSite/content/docs/development/one-api.md
+++ b/docSite/content/docs/development/one-api.md
@ -112,6 +112,7 @@ CHAT_API_KEY=sk-xxxxxx
    {
      "model": "ERNIE-Bot", // 这里的模型需要对应 One API 的模型
      "name": "文心一言", // 对外展示的名称
+      "avatar": "/imgs/model/openai.svg", // 模型的logo
      "maxContext": 16000, // 最大上下文
      "maxResponse": 4000, // 最大回复
      "quoteMaxToken": 13000, // 最大引用内容
@ -135,4 +136,11 @@ CHAT_API_KEY=sk-xxxxxx
 ],
 ```

-添加完后，重启 FastGPT 即可在选择文心一言模型进行对话。**添加向量模型也是类似操作，增加到 `vectorModels`里。**
+### 3. 重启 FastGPT
+
+```bash
+docker-compose down
+docker-compose up -d
+```
+
+重启 FastGPT 即可在选择文心一言模型进行对话。**添加向量模型也是类似操作，增加到 `vectorModels`里。**
--- a/docSite/content/docs/development/upgrading/47.md
+++ b/docSite/content/docs/development/upgrading/47.md
@ -9,11 +9,31 @@ weight: 826

 ## 修改配置文件

-增加一些 Boolean 值，用于决定不同功能块可以使用哪些模型：[点击查看最新的配置文件](/docs/development/configuration/)
+增加一些 Boolean 值，用于决定不同功能块可以使用哪些模型，同时增加了模型的 logo：[点击查看最新的配置文件](/docs/development/configuration/)

+## 初始化脚本
+
+从任意终端，发起 1 个 HTTP 请求。其中 {{rootkey}} 替换成环境变量里的 `rootkey`；{{host}} 替换成自己域名
+
+```bash
+curl --location --request POST 'https://{{host}}/api/admin/initv47' \
+--header 'rootkey: {{rootkey}}' \
+--header 'Content-Type: application/json'
+```
+
+脚本功能：
+1. 初始化插件的 parentId

 ## V4.7 更新说明

 1. 新增 - 工具调用模块，可以让LLM模型根据用户意图，动态的选择其他模型或插件执行。
 2. 新增 - 分类和内容提取支持 functionCall 模式。部分模型支持 functionCall 不支持 ToolCall，也可以使用了。需要把 LLM 模型配置文件里的 `functionCall` 设置为 `true`， `toolChoice`设置为 `false`。如果 `toolChoice` 为 true，会走 tool 模式。
-3. 优化 - 高级编排性能
+3. 新增 - HTTP插件，可实现OpenAPI快速生成插件。
+4. 优化 - 高级编排性能。
+5. 优化 - 抽离 Flow controller 到 packages。
+6. 优化 - AI模型选择。
+7. 修复 - 开源版重排选不上。
+8. 修复 - http 请求 body，不使用时，传入undefined。（会造成部分GET请求失败）
+9. 新增 - 支持 http url 使用变量。
+10. 修复 - 469 的提取的提示词容易造成幻觉。
+11. 修复 - PG HNSW索引未实际生效问题，本次更新后，搜索速度大幅度提升(但是可能会出现精度损失，如果出现精度损失需要参考PgVector文档，对索引进行调整)。详细见：https://github.com/pgvector/pgvector?tab=readme-ov-file#troubleshooting
--- a/docSite/content/docs/use-cases/feishu.md
+++ b/docSite/content/docs/use-cases/feishu.md
@ -1,6 +1,6 @@
 ---
-title: " 接入飞书 "
-description: "FastGPT 接入飞书机器人 "
+title: " 接入飞书(社区文章)"
+description: "FastGPT 接入飞书机器人"
 icon: "chat"
 draft: false
 toc: true
--- a/docSite/content/docs/use-cases/gapier.md
+++ b/docSite/content/docs/use-cases/gapier.md
@ -0,0 +1,60 @@
+---
+title: "使用 Gapier 快速导入Agent工具"
+description: "FastGPT 使用 Gapier 快速导入Agent工具"
+icon: "build"
+draft: false
+toc: true
+weight: 501
+---
+
+FastGPT V4.7版本加入了工具调用，可以兼容 GPTs 的 Actions。这意味着，你可以直接导入兼容 GPTs 的 Agent 工具。
+
+Gapier 是一个在线 GPTs Actions工具，提供了50多种现成工具，并且每天有免费额度进行测试，方便用户试用，官方地址为：[https://gapier.com/](https://gapier.com/)。
+
+![](/imgs/gapierToolResult1.png)
+
+现在，我们开始把 Gapier 的工具导入到 FastGPT 中。
+
+## 1. 创建插件
+
+| Step1 | Step2 | Step3 |
+| --- | --- | --- |
+| ![](/imgs/gapierTool1.png) | ![](/imgs/gapierTool2.png) | 登录[Gapier](https://gapier.com/) 复制相关参数 <br> ![](/imgs/gapierTool3.png) |
+| Step4 | Step5 | Step6 |
+| 自定义请求头: Authorization<br>请求值: Bearer 复制的key <br> ![](/imgs/gapierTool4.png) | ![](/imgs/gapierTool5.png) | ![](/imgs/gapierTool6.png) |
+
+创建完后，如果需要变更，无需重新创建，只需要修改对应参数即可，会自动做差值比较更新。
+
+![](/imgs/gapierTool7.png)
+
+## 2. 应用绑定工具
+
+### 简易模式
+
+| Step1 | Step2 |
+| --- | --- | --- |
+| ![](/imgs/gapierTool8.png) | ![](/imgs/gapierTool9.png) |
+| Step3 | Step4 |
+| ![](/imgs/gapierTool10.png) | ![](/imgs/gapierTool11.png) |
+
+### 高级编排
+
+| Step1 | Step2 |
+| --- | --- | --- |
+| ![](/imgs/gapierTool12.png) | ![](/imgs/gapierTool13.png) |
+| Step3 | Step4 |
+| ![](/imgs/gapierTool14.png) | ![](/imgs/gapierTool15.png) |
+
+![](/imgs/gapierTool16.png)
+
+## 3. 工具调用说明
+
+### 不同模型的区别
+
+不同模型调用工具采用不同的方法，有些模型支持 toolChoice 和 functionCall 效果会更好。不支持这两种方式的模型通过提示词调用，但是效果不是很好，并且为了保证顺利调用，FastGPT内置的提示词，仅支持每次调用一个工具。
+
+具体哪些模型支持 functionCall 可以官网查看（当然，也需要OneAPI支持），同时需要调整模型配置文件中的对应字段（详细看配置字段说明）。
+
+线上版用户，可以在模型选择时，看到是否支持函数调用的标识。
+
+![](/imgs/gapierTool17.png)
--- a/docSite/content/docs/use-cases/wechat.md
+++ b/docSite/content/docs/use-cases/wechat.md
@ -1,5 +1,5 @@
 ---
-title: " 接入微信和企业微信 "
+title: "接入微信和企业微信 "
 description: "FastGPT 接入微信和企业微信 "
 icon: "chat"
 draft: false
--- a/packages/global/common/error/code/team.ts
+++ b/packages/global/common/error/code/team.ts
@ -15,7 +15,7 @@ export enum TeamErrEnum {
 const teamErr = [
  { statusText: TeamErrEnum.teamOverSize, message: 'error.team.overSize' },
  { statusText: TeamErrEnum.unAuthTeam, message: '无权操作该团队' },
-  { statusText: TeamErrEnum.aiPointsNotEnough, message: 'AI积分已用完~' },
+  { statusText: TeamErrEnum.aiPointsNotEnough, message: '' },
  { statusText: TeamErrEnum.datasetSizeNotEnough, message: '知识库容量不足，请先扩容~' },
  { statusText: TeamErrEnum.datasetAmountNotEnough, message: '知识库数量已达上限~' },
  { statusText: TeamErrEnum.appAmountNotEnough, message: '应用数量已达上限~' },
--- a/packages/global/common/error/utils.ts
+++ b/packages/global/common/error/utils.ts
@ -1,7 +1,7 @@
 import { replaceSensitiveText } from '../string/tools';

 export const getErrText = (err: any, def = '') => {
-  const msg: string = typeof err === 'string' ? err : err?.message || def || '';
+  const msg: string = typeof err === 'string' ? err : err?.message ?? def;
  msg && console.log('error =>', msg);
  return replaceSensitiveText(msg);
 };
--- a/packages/global/common/file/constants.ts
+++ b/packages/global/common/file/constants.ts
@ -1,3 +1,4 @@
+/* mongo fs bucket */
 export enum BucketNameEnum {
  dataset = 'dataset'
 }
@ -7,4 +8,4 @@ export const bucketNameMap = {
  }
 };

-export const FileBaseUrl = '/api/common/file/read';
+export const ReadFileBaseUrl = '/api/common/file/read';
--- a/packages/global/common/file/image/constants.ts
+++ b/packages/global/common/file/image/constants.ts
@ -50,3 +50,7 @@ export const mongoImageTypeMap = {
 export const uniqueImageTypeList = Object.entries(mongoImageTypeMap)
  .filter(([key, value]) => value.unique)
  .map(([key]) => key as `${MongoImageTypeEnum}`);
+
+export const FolderIcon = 'file/fill/folder';
+export const FolderImgUrl = '/imgs/files/folder.svg';
+export const HttpImgUrl = '/imgs/module/http.png';
--- a/packages/global/common/system/constants.ts
+++ b/packages/global/common/system/constants.ts
@ -1,2 +1,3 @@
 export const HUMAN_ICON = `/icon/human.svg`;
 export const LOGO_ICON = `/icon/logo.svg`;
+export const HUGGING_FACE_ICON = `/imgs/model/huggingface.svg`;
--- a/packages/global/core/ai/model.d.ts
+++ b/packages/global/core/ai/model.d.ts
@ -1,6 +1,7 @@
 export type LLMModelItemType = {
  model: string;
  name: string;
+  avatar?: string;
  maxContext: number;
  maxResponse: number;
  quoteMaxToken: number;
@ -31,6 +32,7 @@ export type LLMModelItemType = {
 export type VectorModelItemType = {
  model: string;
  name: string;
+  avatar?: string;
  defaultToken: number;
  charsPointsPrice: number;
  maxToken: number;
--- a/projects/app/src/global/core/prompt/AIChat.ts
+++ b/projects/app/src/global/core/prompt/AIChat.ts
@ -1,4 +1,4 @@
-import { PromptTemplateItem } from '@fastgpt/global/core/ai/type.d';
+import { PromptTemplateItem } from '../type.d';

 export const Prompt_QuoteTemplateList: PromptTemplateItem[] = [
  {
--- a/projects/app/src/global/core/prompt/agent.ts
+++ b/projects/app/src/global/core/prompt/agent.ts
@ -58,5 +58,3 @@ Human："{{question}}"

 类型ID=
 `;
-
-export const Prompt_QuestionGuide = `我不太清楚问你什么问题，请帮我生成 3 个问题，引导我继续提问。问题的长度应小于20个字符，按 JSON 格式返回: ["问题1", "问题2", "问题3"]`;
--- a/packages/global/core/app/api.d.ts
+++ b/packages/global/core/app/api.d.ts
@ -12,16 +12,9 @@ export type CreateAppParams = {
 export interface AppUpdateParams {
  name?: string;
  type?: `${AppTypeEnum}`;
-  simpleTemplateId?: string;
  avatar?: string;
  intro?: string;
  modules?: AppSchema['modules'];
  permission?: AppSchema['permission'];
  teamTags?: AppSchema['teamTags'];
 }
-
-export type FormatForm2ModulesProps = {
-  formData: AppSimpleEditFormType;
-  chatModelMaxToken: number;
-  llmModelList: LLMModelItemType[];
-};
--- a/packages/global/core/app/type.d.ts
+++ b/packages/global/core/app/type.d.ts
@ -1,7 +1,12 @@
-import type { AppTTSConfigType, ModuleItemType, VariableItemType } from '../module/type.d';
+import type {
+  AppTTSConfigType,
+  FlowNodeTemplateType,
+  ModuleItemType,
+  VariableItemType
+} from '../module/type.d';
 import { AppTypeEnum } from './constants';
 import { PermissionTypeEnum } from '../../support/permission/constant';
-import type { AIChatModuleProps, DatasetModuleProps } from '../module/node/type.d';
+import type { DatasetModuleProps } from '../module/node/type.d';
 import { VariableInputEnum } from '../module/constants';
 import { SelectedDatasetType } from '../module/api';
 import { DatasetSearchModeEnum } from '../dataset/constants';
@ -13,7 +18,6 @@ export interface AppSchema {
  tmbId: string;
  name: string;
  type: `${AppTypeEnum}`;
-  simpleTemplateId: string;
  avatar: string;
  intro: string;
  updateTime: number;
@ -37,19 +41,6 @@ export type AppDetailType = AppSchema & {
  canWrite: boolean;
 };

-// export type AppSimpleEditFormType = {
-//   aiSettings: AIChatModuleProps;
-//   dataset: DatasetModuleProps & {
-//     searchEmptyText: string;
-//   };
-//   userGuide: {
-//     welcomeText: string;
-//     variables: VariableItemType[];
-//     questionGuide: boolean;
-//     tts: AppTTSConfigType;
-//   };
-// };
-// Since useform cannot infer enumeration types, all enumeration keys can only be undone manually
 export type AppSimpleEditFormType = {
  // templateId: string;
  aiSettings: {
@ -58,8 +49,7 @@ export type AppSimpleEditFormType = {
    temperature: number;
    maxToken: number;
    isResponseAnswerText: boolean;
-    quoteTemplate?: string | undefined;
-    quotePrompt?: string | undefined;
+    maxHistories: number;
  };
  dataset: {
    datasets: SelectedDatasetType;
@ -67,11 +57,11 @@ export type AppSimpleEditFormType = {
    similarity?: number;
    limit?: number;
    usingReRank?: boolean;
-    searchEmptyText?: string;
    datasetSearchUsingExtensionQuery?: boolean;
    datasetSearchExtensionModel?: string;
    datasetSearchExtensionBg?: string;
  };
+  selectedTools: FlowNodeTemplateType[];
  userGuide: {
    welcomeText: string;
    variables: {
@ -94,34 +84,3 @@ export type AppSimpleEditFormType = {
    };
  };
 };
-
-/* simple mode template*/
-export type AppSimpleEditConfigTemplateType = {
-  id: string;
-  name: string;
-  desc: string;
-  systemForm: {
-    aiSettings?: {
-      model?: boolean;
-      systemPrompt?: boolean;
-      temperature?: boolean;
-      maxToken?: boolean;
-      quoteTemplate?: boolean;
-      quotePrompt?: boolean;
-    };
-    dataset?: {
-      datasets?: boolean;
-      similarity?: boolean;
-      limit?: boolean;
-      searchMode: `${DatasetSearchModeEnum}`;
-      usingReRank: boolean;
-      searchEmptyText?: boolean;
-    };
-    userGuide?: {
-      welcomeText?: boolean;
-      variables?: boolean;
-      questionGuide?: boolean;
-      tts?: boolean;
-    };
-  };
-};
--- a/packages/global/core/app/utils.ts
+++ b/packages/global/core/app/utils.ts
@ -1,6 +1,10 @@
 import type { AppSimpleEditFormType } from '../app/type';
 import { FlowNodeTypeEnum } from '../module/node/constant';
-import { ModuleOutputKeyEnum, ModuleInputKeyEnum } from '../module/constants';
+import {
+  ModuleOutputKeyEnum,
+  ModuleInputKeyEnum,
+  FlowNodeTemplateTypeEnum
+} from '../module/constants';
 import type { FlowNodeInputItemType } from '../module/node/type.d';
 import { getGuideModule, splitGuideModule } from '../module/utils';
 import { ModuleItemType } from '../module/type.d';
@ -13,20 +17,19 @@ export const getDefaultAppForm = (): AppSimpleEditFormType => {
      systemPrompt: '',
      temperature: 0,
      isResponseAnswerText: true,
-      quotePrompt: '',
-      quoteTemplate: '',
+      maxHistories: 6,
      maxToken: 4000
    },
    dataset: {
      datasets: [],
      similarity: 0.4,
      limit: 1500,
-      searchEmptyText: '',
      searchMode: DatasetSearchModeEnum.embedding,
      usingReRank: false,
      datasetSearchUsingExtensionQuery: true,
      datasetSearchExtensionBg: ''
    },
+    selectedTools: [],
    userGuide: {
      welcomeText: '',
      variables: [],
@ -47,7 +50,10 @@ export const appModules2Form = ({ modules }: { modules: ModuleItemType[] }) => {
  };

  modules.forEach((module) => {
-    if (module.flowType === FlowNodeTypeEnum.chatNode) {
+    if (
+      module.flowType === FlowNodeTypeEnum.chatNode ||
+      module.flowType === FlowNodeTypeEnum.tools
+    ) {
      defaultAppForm.aiSettings.model = findInputValueByKey(
        module.inputs,
        ModuleInputKeyEnum.aiModel
@ -64,13 +70,9 @@ export const appModules2Form = ({ modules }: { modules: ModuleItemType[] }) => {
        module.inputs,
        ModuleInputKeyEnum.aiChatMaxToken
      );
-      defaultAppForm.aiSettings.quoteTemplate = findInputValueByKey(
+      defaultAppForm.aiSettings.maxHistories = findInputValueByKey(
        module.inputs,
-        ModuleInputKeyEnum.aiChatQuoteTemplate
-      );
-      defaultAppForm.aiSettings.quotePrompt = findInputValueByKey(
-        module.inputs,
-        ModuleInputKeyEnum.aiChatQuotePrompt
+        ModuleInputKeyEnum.history
      );
    } else if (module.flowType === FlowNodeTypeEnum.datasetSearchNode) {
      defaultAppForm.dataset.datasets = findInputValueByKey(
@ -104,17 +106,6 @@ export const appModules2Form = ({ modules }: { modules: ModuleItemType[] }) => {
        module.inputs,
        ModuleInputKeyEnum.datasetSearchExtensionBg
      );
-
-      // empty text
-      const emptyOutputs =
-        module.outputs.find((item) => item.key === ModuleOutputKeyEnum.datasetIsEmpty)?.targets ||
-        [];
-      const emptyOutput = emptyOutputs[0];
-      if (emptyOutput) {
-        const target = modules.find((item) => item.moduleId === emptyOutput.moduleId);
-        defaultAppForm.dataset.searchEmptyText =
-          target?.inputs?.find((item) => item.key === ModuleInputKeyEnum.answerText)?.value || '';
-      }
    } else if (module.flowType === FlowNodeTypeEnum.userGuide) {
      const { welcomeText, variableModules, questionGuide, ttsConfig } = splitGuideModule(
        getGuideModule(modules)
@ -125,6 +116,18 @@ export const appModules2Form = ({ modules }: { modules: ModuleItemType[] }) => {
        questionGuide: questionGuide,
        tts: ttsConfig
      };
+    } else if (module.flowType === FlowNodeTypeEnum.pluginModule) {
+      defaultAppForm.selectedTools.push({
+        id: module.inputs.find((input) => input.key === ModuleInputKeyEnum.pluginId)?.value || '',
+        name: module.name,
+        avatar: module.avatar,
+        intro: module.intro || '',
+        flowType: module.flowType,
+        showStatus: module.showStatus,
+        inputs: module.inputs,
+        outputs: module.outputs,
+        templateType: FlowNodeTemplateTypeEnum.other
+      });
    }
  });

--- a/packages/global/core/chat/adapt.ts
+++ b/packages/global/core/chat/adapt.ts
@ -83,6 +83,7 @@ export const chats2GPTMessages = ({
        });
      }
    } else {
+      //AI
      item.value.forEach((value) => {
        if (value.type === ChatItemValueTypeEnum.tool && value.tools && reserveTool) {
          const tool_calls: ChatCompletionMessageToolCall[] = [];
--- a/packages/global/core/chat/api.d.ts
+++ b/packages/global/core/chat/api.d.ts
@ -3,6 +3,8 @@ export type UpdateChatFeedbackProps = {
  chatId: string;
  chatItemId: string;
  shareId?: string;
+  teamId?: string;
+  teamToken?: string;
  outLinkUid?: string;
  userBadFeedback?: string;
  userGoodFeedback?: string;
--- a/packages/global/core/chat/type.d.ts
+++ b/packages/global/core/chat/type.d.ts
@ -141,7 +141,7 @@ export type ChatHistoryItemResType = DispatchNodeResponseType & {
 };

 /* One tool run response  */
-export type ToolRunResponseItemType = Record<string, any> | Array;
+export type ToolRunResponseItemType = any;
 /* tool module response */
 export type ToolModuleResponseItemType = {
  id: string;
--- a/packages/global/core/dataset/constants.ts
+++ b/packages/global/core/dataset/constants.ts
@ -154,8 +154,5 @@ export const SearchScoreTypeMap = {
  }
 };

-export const FolderIcon = 'file/fill/folder';
-export const FolderImgUrl = '/imgs/files/folder.svg';
-
 export const CustomCollectionIcon = 'common/linkBlue';
 export const LinkCollectionIcon = 'common/linkBlue';
--- a/packages/global/core/module/api.d.ts
+++ b/packages/global/core/module/api.d.ts
@ -13,3 +13,10 @@ export type HttpQueryType = {
  variables: Record<string, any>;
  [key: string]: any;
 };
+
+/* http node */
+export type HttpParamAndHeaderItemType = {
+  key: string;
+  type: string;
+  value: string;
+};
--- a/packages/global/core/module/constants.ts
+++ b/packages/global/core/module/constants.ts
@ -1,4 +1,4 @@
-export enum ModuleTemplateTypeEnum {
+export enum FlowNodeTemplateTypeEnum {
  userGuide = 'userGuide',
  systemInput = 'systemInput',
  tools = 'tools',
@ -87,7 +87,8 @@ export enum ModuleInputKeyEnum {
  runAppSelectApp = 'app',

  // plugin
-  pluginId = 'pluginId'
+  pluginId = 'pluginId',
+  pluginStart = 'pluginStart'
 }

 export enum ModuleOutputKeyEnum {
@ -117,7 +118,10 @@ export enum ModuleOutputKeyEnum {
  selectedTools = 'selectedTools',

  // http
-  httpRawResponse = 'httpRawResponse'
+  httpRawResponse = 'httpRawResponse',
+
+  // plugin
+  pluginStart = 'pluginStart'
 }

 export enum VariableInputEnum {
--- a/packages/global/core/module/node/constant.ts
+++ b/packages/global/core/module/node/constant.ts
@ -21,10 +21,12 @@ export enum FlowNodeInputTypeEnum {

  // ai model select
  selectLLMModel = 'selectLLMModel',
+  settingLLMModel = 'settingLLMModel',

  // dataset special input
  selectDataset = 'selectDataset',
  selectDatasetParamsModal = 'selectDatasetParamsModal',
+  settingDatasetQuotePrompt = 'settingDatasetQuotePrompt',

  hidden = 'hidden',
  custom = 'custom'
@ -57,7 +59,8 @@ export enum FlowNodeTypeEnum {
  pluginInput = 'pluginInput',
  pluginOutput = 'pluginOutput',
  queryExtension = 'cfr',
-  tools = 'tools'
+  tools = 'tools',
+  stopTool = 'stopTool'

  // abandon
 }
--- a/packages/global/core/module/node/type.d.ts
+++ b/packages/global/core/module/node/type.d.ts
@ -102,6 +102,13 @@ export type EditNodeFieldType = {
 };

 /* ------------- item type --------------- */
+export type SettingAIDataType = {
+  model: string;
+  temperature: number;
+  maxToken: number;
+  isResponseAnswerText?: boolean;
+  maxHistories?: number;
+};
 /* ai chat modules props */
 export type AIChatModuleProps = {
  [ModuleInputKeyEnum.aiModel]: string;
--- a/packages/global/core/module/runtime/type.d.ts
+++ b/packages/global/core/module/runtime/type.d.ts
@ -90,6 +90,7 @@ export type DispatchNodeResponseType = {
  // tool
  toolCallTokens?: number;
  toolDetail?: ChatHistoryItemResType[];
+  toolStop?: boolean;
 };

 export type DispatchNodeResultType<T> = {
--- a/packages/global/core/module/template/constants.ts
+++ b/packages/global/core/module/template/constants.ts
@ -8,7 +8,9 @@ import { ClassifyQuestionModule } from './system/classifyQuestion';
 import { ContextExtractModule } from './system/contextExtract';
 import { HttpModule468 } from './system/http468';
 import { HttpModule } from './system/abandon/http';
+
 import { ToolModule } from './system/tools';
+import { StopToolNode } from './system/stopTool';

 import { RunAppModule } from './system/runApp';
 import { PluginInputModule } from './system/pluginInput';
@ -16,11 +18,11 @@ import { PluginOutputModule } from './system/pluginOutput';
 import { RunPluginModule } from './system/runPlugin';
 import { AiQueryExtension } from './system/queryExtension';

-import type { FlowModuleTemplateType, moduleTemplateListType } from '../../module/type.d';
-import { ModuleTemplateTypeEnum } from '../../module/constants';
+import type { FlowNodeTemplateType, moduleTemplateListType } from '../../module/type.d';
+import { FlowNodeTemplateTypeEnum } from '../../module/constants';

 /* app flow module templates */
-export const appSystemModuleTemplates: FlowModuleTemplateType[] = [
+export const appSystemModuleTemplates: FlowNodeTemplateType[] = [
  UserGuideModule,
  UserInputModule,
  AiChatModule,
@ -29,13 +31,14 @@ export const appSystemModuleTemplates: FlowModuleTemplateType[] = [
  DatasetConcatModule,
  RunAppModule,
  ToolModule,
+  StopToolNode,
  ClassifyQuestionModule,
  ContextExtractModule,
  HttpModule468,
  AiQueryExtension
 ];
 /* plugin flow module templates */
-export const pluginSystemModuleTemplates: FlowModuleTemplateType[] = [
+export const pluginSystemModuleTemplates: FlowNodeTemplateType[] = [
  PluginInputModule,
  PluginOutputModule,
  AiChatModule,
@ -44,6 +47,7 @@ export const pluginSystemModuleTemplates: FlowModuleTemplateType[] = [
  DatasetConcatModule,
  RunAppModule,
  ToolModule,
+  StopToolNode,
  ClassifyQuestionModule,
  ContextExtractModule,
  HttpModule468,
@ -51,7 +55,7 @@ export const pluginSystemModuleTemplates: FlowModuleTemplateType[] = [
 ];

 /* all module */
-export const moduleTemplatesFlat: FlowModuleTemplateType[] = [
+export const moduleTemplatesFlat: FlowNodeTemplateType[] = [
  UserGuideModule,
  UserInputModule,
  AiChatModule,
@ -63,6 +67,7 @@ export const moduleTemplatesFlat: FlowModuleTemplateType[] = [
  HttpModule468,
  HttpModule,
  ToolModule,
+  StopToolNode,
  AiChatModule,
  RunAppModule,
  PluginInputModule,
@ -73,43 +78,43 @@ export const moduleTemplatesFlat: FlowModuleTemplateType[] = [

 export const moduleTemplatesList: moduleTemplateListType = [
  {
-    type: ModuleTemplateTypeEnum.userGuide,
-    label: 'core.module.template.Guide module',
+    type: FlowNodeTemplateTypeEnum.userGuide,
+    label: '',
    list: []
  },
  {
-    type: ModuleTemplateTypeEnum.systemInput,
-    label: 'core.module.template.System input module',
-    list: []
-  },
-  {
-    type: ModuleTemplateTypeEnum.textAnswer,
+    type: FlowNodeTemplateTypeEnum.textAnswer,
    label: 'core.module.template.Response module',
    list: []
  },
  {
-    type: ModuleTemplateTypeEnum.functionCall,
+    type: FlowNodeTemplateTypeEnum.functionCall,
    label: 'core.module.template.Function module',
    list: []
  },
  {
-    type: ModuleTemplateTypeEnum.tools,
+    type: FlowNodeTemplateTypeEnum.tools,
    label: 'core.module.template.Tool module',
    list: []
  },
  {
-    type: ModuleTemplateTypeEnum.externalCall,
+    type: FlowNodeTemplateTypeEnum.externalCall,
    label: 'core.module.template.External module',
    list: []
  },
  {
-    type: ModuleTemplateTypeEnum.personalPlugin,
-    label: 'core.module.template.My plugin module',
+    type: FlowNodeTemplateTypeEnum.personalPlugin,
+    label: '',
    list: []
  },
  {
-    type: ModuleTemplateTypeEnum.other,
+    type: FlowNodeTemplateTypeEnum.other,
    label: '其他',
    list: []
+  },
+  {
+    type: FlowNodeTemplateTypeEnum.systemInput,
+    label: 'core.module.template.System input module',
+    list: []
  }
 ];
--- a/packages/global/core/module/template/input.ts
+++ b/packages/global/core/module/template/input.ts
@ -59,7 +59,7 @@ export const Input_Template_DynamicInput: FlowNodeInputItemType = {
  hideInApp: true
 };

-export const Input_Template_AiModel: FlowNodeInputItemType = {
+export const Input_Template_SelectAIModel: FlowNodeInputItemType = {
  key: ModuleInputKeyEnum.aiModel,
  type: FlowNodeInputTypeEnum.selectLLMModel,
  label: 'core.module.input.label.aiModel',
@ -68,6 +68,15 @@ export const Input_Template_AiModel: FlowNodeInputItemType = {
  showTargetInApp: false,
  showTargetInPlugin: false
 };
+export const Input_Template_SettingAiModel: FlowNodeInputItemType = {
+  key: ModuleInputKeyEnum.aiModel,
+  type: FlowNodeInputTypeEnum.settingLLMModel,
+  label: 'core.module.input.label.aiModel',
+  required: true,
+  valueType: ModuleIOValueTypeEnum.string,
+  showTargetInApp: false,
+  showTargetInPlugin: false
+};

 export const Input_Template_System_Prompt: FlowNodeInputItemType = {
  key: ModuleInputKeyEnum.aiSystemPrompt,
@ -83,7 +92,7 @@ export const Input_Template_System_Prompt: FlowNodeInputItemType = {

 export const Input_Template_Dataset_Quote: FlowNodeInputItemType = {
  key: ModuleInputKeyEnum.aiChatDatasetQuote,
-  type: FlowNodeInputTypeEnum.target,
+  type: FlowNodeInputTypeEnum.settingDatasetQuotePrompt,
  label: '知识库引用',
  description: 'core.module.Dataset quote.Input description',
  valueType: ModuleIOValueTypeEnum.datasetQuote,
--- a/packages/global/core/module/template/system/abandon/http.ts
+++ b/packages/global/core/module/template/system/abandon/http.ts
@ -3,11 +3,11 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../../node/constant';
-import { FlowModuleTemplateType } from '../../../type';
+import { FlowNodeTemplateType } from '../../../type';
 import {
  ModuleIOValueTypeEnum,
  ModuleInputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum
 } from '../../../constants';
 import {
  Input_Template_AddInputParam,
@ -16,9 +16,9 @@ import {
 } from '../../input';
 import { Output_Template_AddOutput, Output_Template_Finish } from '../../output';

-export const HttpModule: FlowModuleTemplateType = {
+export const HttpModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.httpRequest,
-  templateType: ModuleTemplateTypeEnum.externalCall,
+  templateType: FlowNodeTemplateTypeEnum.externalCall,
  flowType: FlowNodeTypeEnum.httpRequest,
  avatar: '/imgs/module/http.png',
  name: 'core.module.template.Http request',
--- a/packages/global/core/module/template/system/aiChat.ts
+++ b/packages/global/core/module/template/system/aiChat.ts
@ -3,15 +3,15 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import { FlowNodeTemplateType } from '../../type.d';
 import {
  ModuleIOValueTypeEnum,
  ModuleInputKeyEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum
 } from '../../constants';
 import {
-  Input_Template_AiModel,
+  Input_Template_SettingAiModel,
  Input_Template_Dataset_Quote,
  Input_Template_History,
  Input_Template_Switch,
@ -21,18 +21,18 @@ import {
 import { chatNodeSystemPromptTip } from '../tip';
 import { Output_Template_Finish, Output_Template_UserChatInput } from '../output';

-export const AiChatModule: FlowModuleTemplateType = {
+export const AiChatModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.chatNode,
-  templateType: ModuleTemplateTypeEnum.textAnswer,
+  templateType: FlowNodeTemplateTypeEnum.textAnswer,
  flowType: FlowNodeTypeEnum.chatNode,
  avatar: '/imgs/module/AI.png',
  name: 'AI 对话',
  intro: 'AI 大模型对话',
  showStatus: true,
-  // isTool: true,
+  isTool: true,
  inputs: [
    Input_Template_Switch,
-    Input_Template_AiModel,
+    Input_Template_SettingAiModel,
    // --- settings modal
    {
      key: ModuleInputKeyEnum.aiChatTemperature,
@ -83,14 +83,6 @@ export const AiChatModule: FlowModuleTemplateType = {
      showTargetInApp: false,
      showTargetInPlugin: false
    },
-    {
-      key: ModuleInputKeyEnum.aiChatSettingModal,
-      type: FlowNodeInputTypeEnum.aiSettings,
-      label: '',
-      valueType: ModuleIOValueTypeEnum.any,
-      showTargetInApp: false,
-      showTargetInPlugin: false
-    },
    // settings modal ---
    {
      ...Input_Template_System_Prompt,
--- a/packages/global/core/module/template/system/assignedAnswer.ts
+++ b/packages/global/core/module/template/system/assignedAnswer.ts
@ -1,12 +1,16 @@
 import { FlowNodeInputTypeEnum, FlowNodeTypeEnum } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
-import { ModuleIOValueTypeEnum, ModuleInputKeyEnum, ModuleTemplateTypeEnum } from '../../constants';
+import { FlowNodeTemplateType } from '../../type.d';
+import {
+  ModuleIOValueTypeEnum,
+  ModuleInputKeyEnum,
+  FlowNodeTemplateTypeEnum
+} from '../../constants';
 import { Input_Template_Switch } from '../input';
 import { Output_Template_Finish } from '../output';

-export const AssignedAnswerModule: FlowModuleTemplateType = {
+export const AssignedAnswerModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.answerNode,
-  templateType: ModuleTemplateTypeEnum.textAnswer,
+  templateType: FlowNodeTemplateTypeEnum.textAnswer,
  flowType: FlowNodeTypeEnum.answerNode,
  avatar: '/imgs/module/reply.png',
  name: '指定回复',
--- a/packages/global/core/module/template/system/classifyQuestion.ts
+++ b/packages/global/core/module/template/system/classifyQuestion.ts
@ -3,10 +3,14 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
-import { ModuleIOValueTypeEnum, ModuleInputKeyEnum, ModuleTemplateTypeEnum } from '../../constants';
+import { FlowNodeTemplateType } from '../../type.d';
 import {
-  Input_Template_AiModel,
+  ModuleIOValueTypeEnum,
+  ModuleInputKeyEnum,
+  FlowNodeTemplateTypeEnum
+} from '../../constants';
+import {
+  Input_Template_SelectAIModel,
  Input_Template_History,
  Input_Template_Switch,
  Input_Template_UserChatInput
@ -15,9 +19,9 @@ import { Output_Template_UserChatInput } from '../output';
 import { Input_Template_System_Prompt } from '../input';
 import { LLMModelTypeEnum } from '../../../ai/constants';

-export const ClassifyQuestionModule: FlowModuleTemplateType = {
+export const ClassifyQuestionModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.classifyQuestion,
-  templateType: ModuleTemplateTypeEnum.functionCall,
+  templateType: FlowNodeTemplateTypeEnum.functionCall,
  flowType: FlowNodeTypeEnum.classifyQuestion,
  avatar: '/imgs/module/cq.png',
  name: '问题分类',
@ -26,7 +30,7 @@ export const ClassifyQuestionModule: FlowModuleTemplateType = {
  inputs: [
    Input_Template_Switch,
    {
-      ...Input_Template_AiModel,
+      ...Input_Template_SelectAIModel,
      llmModelType: LLMModelTypeEnum.classify
    },
    {
--- a/packages/global/core/module/template/system/contextExtract.ts
+++ b/packages/global/core/module/template/system/contextExtract.ts
@ -3,19 +3,23 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import { FlowNodeTemplateType } from '../../type.d';
 import {
  ModuleIOValueTypeEnum,
  ModuleInputKeyEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum
 } from '../../constants';
-import { Input_Template_AiModel, Input_Template_History, Input_Template_Switch } from '../input';
+import {
+  Input_Template_SelectAIModel,
+  Input_Template_History,
+  Input_Template_Switch
+} from '../input';
 import { LLMModelTypeEnum } from '../../../ai/constants';

-export const ContextExtractModule: FlowModuleTemplateType = {
+export const ContextExtractModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.contentExtract,
-  templateType: ModuleTemplateTypeEnum.functionCall,
+  templateType: FlowNodeTemplateTypeEnum.functionCall,
  flowType: FlowNodeTypeEnum.contentExtract,
  avatar: '/imgs/module/extract.png',
  name: '文本内容提取',
@ -25,7 +29,7 @@ export const ContextExtractModule: FlowModuleTemplateType = {
  inputs: [
    Input_Template_Switch,
    {
-      ...Input_Template_AiModel,
+      ...Input_Template_SelectAIModel,
      llmModelType: LLMModelTypeEnum.extractFields
    },
    {
@ -35,7 +39,6 @@ export const ContextExtractModule: FlowModuleTemplateType = {
      label: '提取要求描述',
      description:
        '给AI一些对应的背景知识或要求描述，引导AI更好的完成任务。\n该输入框可使用全局变量。',
-      required: true,
      placeholder:
        '例如: \n1. 当前时间为: {{cTime}}。你是一个实验室预约助手，你的任务是帮助用户预约实验室，从文本中获取对应的预约信息。\n2. 你是谷歌搜索助手，需要从文本中提取出合适的搜索词。',
      showTargetInApp: true,
--- a/packages/global/core/module/template/system/datasetConcat.ts
+++ b/packages/global/core/module/template/system/datasetConcat.ts
@ -3,12 +3,12 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import { FlowNodeTemplateType } from '../../type.d';
 import {
  ModuleIOValueTypeEnum,
  ModuleInputKeyEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum
 } from '../../constants';
 import { Input_Template_Dataset_Quote, Input_Template_Switch } from '../input';
 import { Output_Template_Finish } from '../output';
@ -20,10 +20,10 @@ export const getOneQuoteInputTemplate = (key = getNanoid()) => ({
  type: FlowNodeInputTypeEnum.hidden
 });

-export const DatasetConcatModule: FlowModuleTemplateType = {
+export const DatasetConcatModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.datasetConcatNode,
  flowType: FlowNodeTypeEnum.datasetConcatNode,
-  templateType: ModuleTemplateTypeEnum.tools,
+  templateType: FlowNodeTemplateTypeEnum.other,
  avatar: '/imgs/module/concat.svg',
  name: '知识库搜索引用合并',
  intro: '可以将多个知识库搜索结果进行合并输出。使用 RRF 的合并方式进行最终排序输出。',
--- a/packages/global/core/module/template/system/datasetSearch.ts
+++ b/packages/global/core/module/template/system/datasetSearch.ts
@ -3,20 +3,20 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import { FlowNodeTemplateType } from '../../type.d';
 import {
  ModuleIOValueTypeEnum,
  ModuleInputKeyEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum
 } from '../../constants';
 import { Input_Template_Switch, Input_Template_UserChatInput } from '../input';
 import { Output_Template_Finish, Output_Template_UserChatInput } from '../output';
 import { DatasetSearchModeEnum } from '../../../dataset/constants';

-export const DatasetSearchModule: FlowModuleTemplateType = {
+export const DatasetSearchModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.datasetSearchNode,
-  templateType: ModuleTemplateTypeEnum.functionCall,
+  templateType: FlowNodeTemplateTypeEnum.functionCall,
  flowType: FlowNodeTypeEnum.datasetSearchNode,
  avatar: '/imgs/module/db.png',
  name: '知识库搜索',
--- a/packages/global/core/module/template/system/http468.ts
+++ b/packages/global/core/module/template/system/http468.ts
@ -3,12 +3,12 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type';
+import { FlowNodeTemplateType } from '../../type';
 import {
  ModuleIOValueTypeEnum,
  ModuleInputKeyEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum
 } from '../../constants';
 import {
  Input_Template_AddInputParam,
@ -17,9 +17,9 @@ import {
 } from '../input';
 import { Output_Template_AddOutput, Output_Template_Finish } from '../output';

-export const HttpModule468: FlowModuleTemplateType = {
+export const HttpModule468: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.httpRequest468,
-  templateType: ModuleTemplateTypeEnum.externalCall,
+  templateType: FlowNodeTemplateTypeEnum.externalCall,
  flowType: FlowNodeTypeEnum.httpRequest468,
  avatar: '/imgs/module/http.png',
  name: 'HTTP 请求',
--- a/packages/global/core/module/template/system/pluginInput.ts
+++ b/packages/global/core/module/template/system/pluginInput.ts
@ -1,15 +1,43 @@
-import { ModuleTemplateTypeEnum } from '../../constants';
-import { FlowNodeTypeEnum } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import {
+  FlowNodeTemplateTypeEnum,
+  ModuleIOValueTypeEnum,
+  ModuleInputKeyEnum,
+  ModuleOutputKeyEnum
+} from '../../constants';
+import {
+  FlowNodeInputTypeEnum,
+  FlowNodeOutputTypeEnum,
+  FlowNodeTypeEnum
+} from '../../node/constant';
+import { FlowNodeTemplateType } from '../../type.d';

-export const PluginInputModule: FlowModuleTemplateType = {
+export const PluginInputModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.pluginInput,
-  templateType: ModuleTemplateTypeEnum.systemInput,
+  templateType: FlowNodeTemplateTypeEnum.systemInput,
  flowType: FlowNodeTypeEnum.pluginInput,
  avatar: '/imgs/module/input.png',
  name: '定义插件输入',
  intro: '自定义配置外部输入，使用插件时，仅暴露自定义配置的输入',
  showStatus: false,
-  inputs: [],
-  outputs: []
+  inputs: [
+    {
+      key: ModuleInputKeyEnum.pluginStart,
+      type: FlowNodeInputTypeEnum.hidden,
+      valueType: ModuleIOValueTypeEnum.boolean,
+      label: '插件开始运行',
+      description:
+        '插件开始运行时，会输出一个 True 的标识。有时候，插件不会有额外的的输入，为了顺利的进入下一个阶段，你可以将该值连接到下一个节点的触发器中。',
+      showTargetInApp: true,
+      showTargetInPlugin: true
+    }
+  ],
+  outputs: [
+    {
+      key: ModuleOutputKeyEnum.pluginStart,
+      label: '插件开始运行',
+      type: FlowNodeOutputTypeEnum.source,
+      valueType: ModuleIOValueTypeEnum.boolean,
+      targets: []
+    }
+  ]
 };
--- a/packages/global/core/module/template/system/pluginOutput.ts
+++ b/packages/global/core/module/template/system/pluginOutput.ts
@ -1,10 +1,10 @@
-import { ModuleTemplateTypeEnum } from '../../constants';
+import { FlowNodeTemplateTypeEnum } from '../../constants';
 import { FlowNodeTypeEnum } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import { FlowNodeTemplateType } from '../../type.d';

-export const PluginOutputModule: FlowModuleTemplateType = {
+export const PluginOutputModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.pluginOutput,
-  templateType: ModuleTemplateTypeEnum.systemInput,
+  templateType: FlowNodeTemplateTypeEnum.systemInput,
  flowType: FlowNodeTypeEnum.pluginOutput,
  avatar: '/imgs/module/output.png',
  name: '定义插件输出',
--- a/packages/global/core/module/template/system/queryExtension.ts
+++ b/packages/global/core/module/template/system/queryExtension.ts
@ -3,25 +3,25 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type';
+import { FlowNodeTemplateType } from '../../type';
 import {
  ModuleIOValueTypeEnum,
  ModuleInputKeyEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum
 } from '../../constants';
 import {
  Input_Template_History,
  Input_Template_Switch,
  Input_Template_UserChatInput,
-  Input_Template_AiModel
+  Input_Template_SelectAIModel
 } from '../input';
 import { Output_Template_UserChatInput } from '../output';
 import { LLMModelTypeEnum } from '../../../ai/constants';

-export const AiQueryExtension: FlowModuleTemplateType = {
+export const AiQueryExtension: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.chatNode,
-  templateType: ModuleTemplateTypeEnum.other,
+  templateType: FlowNodeTemplateTypeEnum.other,
  flowType: FlowNodeTypeEnum.queryExtension,
  avatar: '/imgs/module/cfr.svg',
  name: '问题优化',
@ -31,7 +31,7 @@ export const AiQueryExtension: FlowModuleTemplateType = {
  inputs: [
    Input_Template_Switch,
    {
-      ...Input_Template_AiModel,
+      ...Input_Template_SelectAIModel,
      llmModelType: LLMModelTypeEnum.queryExtension
    },
    {
--- a/packages/global/core/module/template/system/runApp.ts
+++ b/packages/global/core/module/template/system/runApp.ts
@ -3,12 +3,12 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import { FlowNodeTemplateType } from '../../type.d';
 import {
  ModuleIOValueTypeEnum,
  ModuleInputKeyEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum
 } from '../../constants';
 import {
  Input_Template_History,
@ -17,9 +17,9 @@ import {
 } from '../input';
 import { Output_Template_Finish, Output_Template_UserChatInput } from '../output';

-export const RunAppModule: FlowModuleTemplateType = {
+export const RunAppModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.runApp,
-  templateType: ModuleTemplateTypeEnum.externalCall,
+  templateType: FlowNodeTemplateTypeEnum.externalCall,
  flowType: FlowNodeTypeEnum.runApp,
  avatar: '/imgs/module/app.png',
  name: '应用调用',
--- a/packages/global/core/module/template/system/runPlugin.ts
+++ b/packages/global/core/module/template/system/runPlugin.ts
@ -1,10 +1,10 @@
-import { ModuleTemplateTypeEnum } from '../../constants';
+import { FlowNodeTemplateTypeEnum } from '../../constants';
 import { FlowNodeTypeEnum } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import { FlowNodeTemplateType } from '../../type.d';

-export const RunPluginModule: FlowModuleTemplateType = {
+export const RunPluginModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.pluginModule,
-  templateType: ModuleTemplateTypeEnum.externalCall,
+  templateType: FlowNodeTemplateTypeEnum.externalCall,
  flowType: FlowNodeTypeEnum.pluginModule,
  intro: '',
  name: '',
--- a/packages/global/core/module/template/system/stopTool.ts
+++ b/packages/global/core/module/template/system/stopTool.ts
@ -0,0 +1,16 @@
+import { FlowNodeTypeEnum } from '../../node/constant';
+import { FlowNodeTemplateType } from '../../type.d';
+import { FlowNodeTemplateTypeEnum } from '../../constants';
+import { Input_Template_Switch } from '../input';
+
+export const StopToolNode: FlowNodeTemplateType = {
+  id: FlowNodeTypeEnum.stopTool,
+  templateType: FlowNodeTemplateTypeEnum.functionCall,
+  flowType: FlowNodeTypeEnum.stopTool,
+  avatar: '/imgs/module/toolStop.svg',
+  name: '工具调用终止',
+  intro:
+    '该模块需配置工具调用使用。当该模块被执行时，本次工具调用将会强制结束，并且不再调用AI针对工具调用结果回答问题。',
+  inputs: [Input_Template_Switch],
+  outputs: []
+};
--- a/packages/global/core/module/template/system/tools.ts
+++ b/packages/global/core/module/template/system/tools.ts
@ -1,12 +1,17 @@
-import { FlowNodeOutputTypeEnum, FlowNodeTypeEnum } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import {
+  FlowNodeInputTypeEnum,
+  FlowNodeOutputTypeEnum,
+  FlowNodeTypeEnum
+} from '../../node/constant';
+import { FlowNodeTemplateType } from '../../type.d';
 import {
  ModuleIOValueTypeEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum,
+  ModuleInputKeyEnum
 } from '../../constants';
 import {
-  Input_Template_AiModel,
+  Input_Template_SettingAiModel,
  Input_Template_History,
  Input_Template_Switch,
  Input_Template_System_Prompt,
@ -16,19 +21,43 @@ import { chatNodeSystemPromptTip } from '../tip';
 import { Output_Template_Finish, Output_Template_UserChatInput } from '../output';
 import { LLMModelTypeEnum } from '../../../ai/constants';

-export const ToolModule: FlowModuleTemplateType = {
+export const ToolModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.tools,
  flowType: FlowNodeTypeEnum.tools,
-  templateType: ModuleTemplateTypeEnum.functionCall,
+  templateType: FlowNodeTemplateTypeEnum.functionCall,
  avatar: '/imgs/module/tool.svg',
  name: '工具调用（实验）',
-  intro: '通过AI模型自动选择一个或多个工具进行调用。工具可以是其他功能块或插件。',
+  intro: '通过AI模型自动选择一个或多个功能块进行调用，也可以对插件进行调用。',
  showStatus: true,
  inputs: [
    Input_Template_Switch,
    {
-      ...Input_Template_AiModel,
-      llmModelType: LLMModelTypeEnum.toolCall
+      ...Input_Template_SettingAiModel,
+      llmModelType: LLMModelTypeEnum.all
+    },
+    {
+      key: ModuleInputKeyEnum.aiChatTemperature,
+      type: FlowNodeInputTypeEnum.hidden, // Set in the pop-up window
+      label: '',
+      value: 0,
+      valueType: ModuleIOValueTypeEnum.number,
+      min: 0,
+      max: 10,
+      step: 1,
+      showTargetInApp: false,
+      showTargetInPlugin: false
+    },
+    {
+      key: ModuleInputKeyEnum.aiChatMaxToken,
+      type: FlowNodeInputTypeEnum.hidden, // Set in the pop-up window
+      label: '',
+      value: 2000,
+      valueType: ModuleIOValueTypeEnum.number,
+      min: 100,
+      max: 4000,
+      step: 50,
+      showTargetInApp: false,
+      showTargetInPlugin: false
    },
    {
      ...Input_Template_System_Prompt,
--- a/packages/global/core/module/template/system/userGuide.ts
+++ b/packages/global/core/module/template/system/userGuide.ts
@ -1,11 +1,15 @@
 import { FlowNodeInputTypeEnum, FlowNodeTypeEnum } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import { FlowNodeTemplateType } from '../../type.d';
 import { userGuideTip } from '../tip';
-import { ModuleIOValueTypeEnum, ModuleInputKeyEnum, ModuleTemplateTypeEnum } from '../../constants';
+import {
+  ModuleIOValueTypeEnum,
+  ModuleInputKeyEnum,
+  FlowNodeTemplateTypeEnum
+} from '../../constants';

-export const UserGuideModule: FlowModuleTemplateType = {
+export const UserGuideModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.userGuide,
-  templateType: ModuleTemplateTypeEnum.userGuide,
+  templateType: FlowNodeTemplateTypeEnum.userGuide,
  flowType: FlowNodeTypeEnum.userGuide,
  avatar: '/imgs/module/userGuide.png',
  name: '全局配置',
--- a/packages/global/core/module/template/system/userInput.ts
+++ b/packages/global/core/module/template/system/userInput.ts
@ -3,17 +3,17 @@ import {
  FlowNodeOutputTypeEnum,
  FlowNodeTypeEnum
 } from '../../node/constant';
-import { FlowModuleTemplateType } from '../../type.d';
+import { FlowNodeTemplateType } from '../../type.d';
 import {
  ModuleIOValueTypeEnum,
  ModuleInputKeyEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum
+  FlowNodeTemplateTypeEnum
 } from '../../constants';

-export const UserInputModule: FlowModuleTemplateType = {
+export const UserInputModule: FlowNodeTemplateType = {
  id: FlowNodeTypeEnum.questionInput,
-  templateType: ModuleTemplateTypeEnum.systemInput,
+  templateType: FlowNodeTemplateTypeEnum.systemInput,
  flowType: FlowNodeTypeEnum.questionInput,
  avatar: '/imgs/module/userChatInput.svg',
  name: '对话入口',
--- a/packages/global/core/module/type.d.ts
+++ b/packages/global/core/module/type.d.ts
@ -2,7 +2,7 @@ import { FlowNodeTypeEnum } from './node/constant';
 import {
  ModuleIOValueTypeEnum,
  ModuleOutputKeyEnum,
-  ModuleTemplateTypeEnum,
+  FlowNodeTemplateTypeEnum,
  VariableInputEnum
 } from './constants';
 import { DispatchNodeResponseKeyEnum } from './runtime/constants';
@ -15,10 +15,11 @@ import {
 } from '../chat/type';
 import { ChatNodeUsageType } from '../../support/wallet/bill/type';
 import { RunningModuleItemType } from './runtime/type';
+import { PluginTypeEnum } from 'core/plugin/constants';

-export type FlowModuleTemplateType = {
+export type FlowNodeTemplateType = {
  id: string; // module id, unique
-  templateType: `${ModuleTemplateTypeEnum}`;
+  templateType: `${FlowNodeTemplateTypeEnum}`;
  flowType: `${FlowNodeTypeEnum}`; // render node card
  avatar?: string;
  name: string;
@ -27,14 +28,18 @@ export type FlowModuleTemplateType = {
  showStatus?: boolean; // chatting response step status
  inputs: FlowNodeInputItemType[];
  outputs: FlowNodeOutputItemType[];
+
+  // plugin data
+  pluginType?: `${PluginTypeEnum}`;
+  parentId?: string;
 };
-export type FlowModuleItemType = FlowModuleTemplateType & {
+export type FlowModuleItemType = FlowNodeTemplateType & {
  moduleId: string;
 };
 export type moduleTemplateListType = {
-  type: `${ModuleTemplateTypeEnum}`;
+  type: `${FlowNodeTemplateTypeEnum}`;
  label: string;
-  list: FlowModuleTemplateType[];
+  list: FlowNodeTemplateType[];
 }[];

 // store module type
--- a/packages/global/core/module/utils.ts
+++ b/packages/global/core/module/utils.ts
@ -9,6 +9,7 @@ import { FlowNodeInputItemType, FlowNodeOutputItemType } from './node/type';
 import { AppTTSConfigType, ModuleItemType, VariableItemType } from './type';
 import { Input_Template_Switch } from './template/input';
 import { EditorVariablePickerType } from '../../../web/components/common/Textarea/PromptEditor/type';
+import { Output_Template_Finish } from './template/output';

 /* module  */
 export const getGuideModule = (modules: ModuleItemType[]) =>
@ -92,13 +93,16 @@ export const plugin2ModuleIO = (
            connected: false
          }))
        ]
-      : [],
+      : [Input_Template_Switch],
    outputs: pluginOutput
-      ? pluginOutput.outputs.map((item) => ({
-          ...item,
-          edit: false
-        }))
-      : []
+      ? [
+          ...pluginOutput.outputs.map((item) => ({
+            ...item,
+            edit: false
+          })),
+          Output_Template_Finish
+        ]
+      : [Output_Template_Finish]
  };
 };

--- a/packages/global/core/plugin/constants.ts
+++ b/packages/global/core/plugin/constants.ts
@ -27,6 +27,26 @@ export const defaultModules: ModuleItemType[] = [
  }
 ];

+export enum PluginTypeEnum {
+  folder = 'folder',
+  custom = 'custom',
+  http = 'http'
+}
+export const pluginTypeMap = {
+  [PluginTypeEnum.folder]: {
+    label: '文件夹',
+    icon: 'file/fill/folder'
+  },
+  [PluginTypeEnum.custom]: {
+    label: '自定义',
+    icon: 'common/custom'
+  },
+  [PluginTypeEnum.http]: {
+    label: 'HTTP',
+    icon: 'common/http'
+  }
+};
+
 export enum PluginSourceEnum {
  personal = 'personal',
  community = 'community',
--- a/packages/global/core/plugin/controller.d.ts
+++ b/packages/global/core/plugin/controller.d.ts
@ -1,21 +1,40 @@
 import type { ModuleItemType } from '../module/type.d';
+import { PluginTypeEnum } from './constants';
+import { HttpAuthMethodType } from './httpPlugin/type';

 export type CreateOnePluginParams = {
  name: string;
  avatar: string;
  intro: string;
-  modules?: ModuleItemType[];
+  modules: ModuleItemType[];
+  parentId: string | null;
+  type: `${PluginTypeEnum}`;
+  metadata?: {
+    apiSchemaStr?: string;
+    customHeaders?: string;
+  };
 };
 export type UpdatePluginParams = {
  id: string;
+  parentId?: string | null;
  name?: string;
  avatar?: string;
  intro?: string;
  modules?: ModuleItemType[];
+  metadata?: {
+    apiSchemaStr?: string;
+    customHeaders?: string;
+  };
 };
 export type PluginListItemType = {
  _id: string;
+  parentId: string;
+  type: `${PluginTypeEnum}`;
  name: string;
  avatar: string;
  intro: string;
+  metadata?: {
+    apiSchemaStr?: string;
+    customHeaders?: string;
+  };
 };
--- a/packages/global/core/plugin/httpPlugin/type.d.ts
+++ b/packages/global/core/plugin/httpPlugin/type.d.ts
@ -0,0 +1,13 @@
+export type PathDataType = {
+  name: string;
+  description: string;
+  method: string;
+  path: string;
+  params: any[];
+  request: any;
+};
+
+export type OpenApiJsonSchema = {
+  pathData: PathDataType[];
+  serverPath: string;
+};
--- a/packages/global/core/plugin/httpPlugin/utils.ts
+++ b/packages/global/core/plugin/httpPlugin/utils.ts
@ -0,0 +1,516 @@
+import { getNanoid } from '../../../common/string/tools';
+import { OpenApiJsonSchema } from './type';
+import yaml from 'js-yaml';
+import { OpenAPIV3 } from 'openapi-types';
+import { PluginTypeEnum } from '../constants';
+import { FlowNodeInputItemType, FlowNodeOutputItemType } from '../../module/node/type';
+import { FlowNodeInputTypeEnum, FlowNodeOutputTypeEnum } from '../../module/node/constant';
+import { ModuleIOValueTypeEnum } from '../../module/constants';
+import { PluginInputModule } from '../../module/template/system/pluginInput';
+import { PluginOutputModule } from '../../module/template/system/pluginOutput';
+import { HttpModule468 } from '../../module/template/system/http468';
+import { HttpParamAndHeaderItemType } from '../../module/api';
+import { CreateOnePluginParams } from '../controller';
+import { ModuleItemType } from '../../module/type';
+import { HttpImgUrl } from '../../../common/file/image/constants';
+
+export const str2OpenApiSchema = (yamlStr = ''): OpenApiJsonSchema => {
+  try {
+    const data: OpenAPIV3.Document = (() => {
+      try {
+        return JSON.parse(yamlStr);
+      } catch (jsonError) {
+        return yaml.load(yamlStr, { schema: yaml.FAILSAFE_SCHEMA });
+      }
+    })();
+
+    const serverPath = data.servers?.[0].url || '';
+    const pathData = Object.keys(data.paths)
+      .map((path) => {
+        const methodData: any = data.paths[path];
+        return Object.keys(methodData)
+          .filter((method) =>
+            ['get', 'post', 'put', 'delete', 'patch'].includes(method.toLocaleLowerCase())
+          )
+          .map((method) => {
+            const methodInfo = methodData[method];
+            if (methodInfo.deprecated) return;
+            const result = {
+              path,
+              method,
+              name: methodInfo.operationId || path,
+              description: methodInfo.description,
+              params: methodInfo.parameters,
+              request: methodInfo?.requestBody
+            };
+            return result;
+          });
+      })
+      .flat()
+      .filter(Boolean) as OpenApiJsonSchema['pathData'];
+
+    return { pathData, serverPath };
+  } catch (err) {
+    throw new Error('Invalid Schema');
+  }
+};
+
+export const httpApiSchema2Plugins = ({
+  parentId,
+  apiSchemaStr = '',
+  customHeader = ''
+}: {
+  parentId: string;
+  apiSchemaStr?: string;
+  customHeader?: string;
+}): CreateOnePluginParams[] => {
+  const jsonSchema = str2OpenApiSchema(apiSchemaStr);
+  const baseUrl = jsonSchema.serverPath;
+
+  return jsonSchema.pathData.map((item) => {
+    const pluginOutputId = getNanoid();
+    const httpId = getNanoid();
+    const pluginOutputKey = 'result';
+
+    const properties = item.request?.content?.['application/json']?.schema?.properties;
+    const propsKeys = properties ? Object.keys(properties) : [];
+
+    const pluginInputs: FlowNodeInputItemType[] = [
+      ...(item.params?.map((param: any) => {
+        return {
+          key: param.name,
+          valueType: ModuleIOValueTypeEnum.string,
+          label: param.name,
+          type: FlowNodeInputTypeEnum.target,
+          required: param.required,
+          description: param.description,
+          edit: true,
+          editField: {
+            key: true,
+            name: true,
+            description: true,
+            required: true,
+            dataType: true,
+            inputType: true,
+            isToolInput: true
+          },
+          connected: true,
+          toolDescription: param.description
+        };
+      }) || []),
+      ...(propsKeys?.map((key) => {
+        const prop = properties[key];
+        return {
+          key,
+          valueType: ModuleIOValueTypeEnum.string,
+          label: key,
+          type: FlowNodeInputTypeEnum.target,
+          required: false,
+          description: prop.description,
+          edit: true,
+          editField: {
+            key: true,
+            name: true,
+            description: true,
+            required: true,
+            dataType: true,
+            inputType: true,
+            isToolInput: true
+          },
+          connected: true,
+          toolDescription: prop.description
+        };
+      }) || [])
+    ];
+
+    const pluginOutputs: FlowNodeOutputItemType[] = [
+      ...(item.params?.map((param: any) => {
+        return {
+          key: param.name,
+          valueType: ModuleIOValueTypeEnum.string,
+          label: param.name,
+          type: FlowNodeOutputTypeEnum.source,
+          edit: true,
+          targets: [
+            {
+              moduleId: httpId,
+              key: param.name
+            }
+          ]
+        };
+      }) || []),
+      ...(propsKeys?.map((key) => {
+        return {
+          key,
+          valueType: ModuleIOValueTypeEnum.string,
+          label: key,
+          type: FlowNodeOutputTypeEnum.source,
+          edit: true,
+          targets: [
+            {
+              moduleId: httpId,
+              key
+            }
+          ]
+        };
+      }) || [])
+    ];
+
+    const httpInputs: FlowNodeInputItemType[] = [
+      ...(item.params?.map((param: any) => {
+        return {
+          key: param.name,
+          valueType: ModuleIOValueTypeEnum.string,
+          label: param.name,
+          type: FlowNodeInputTypeEnum.target,
+          description: param.description,
+          edit: true,
+          editField: {
+            key: true,
+            description: true,
+            dataType: true
+          },
+          connected: true
+        };
+      }) || []),
+      ...(propsKeys?.map((key) => {
+        const prop = properties[key];
+        return {
+          key,
+          valueType: ModuleIOValueTypeEnum.string,
+          label: key,
+          type: FlowNodeInputTypeEnum.target,
+          description: prop.description,
+          edit: true,
+          editField: {
+            key: true,
+            description: true,
+            dataType: true
+          },
+          connected: true
+        };
+      }) || [])
+    ];
+
+    /* http node setting */
+    const httpNodeParams: HttpParamAndHeaderItemType[] = [];
+    const httpNodeHeaders: HttpParamAndHeaderItemType[] = [];
+    let httpNodeBody = '{}';
+    const requestUrl = `${baseUrl}${item.path}`;
+
+    if (item.params && item.params.length > 0) {
+      for (const param of item.params) {
+        if (param.in === 'header') {
+          httpNodeHeaders.push({
+            key: param.name,
+            type: param.schema?.type || ModuleIOValueTypeEnum.string,
+            value: `{{${param.name}}}`
+          });
+        } else if (param.in === 'body') {
+          httpNodeBody = JSON.stringify(
+            { ...JSON.parse(httpNodeBody), [param.name]: `{{${param.name}}}` },
+            null,
+            2
+          );
+        } else if (param.in === 'query') {
+          httpNodeParams.push({
+            key: param.name,
+            type: param.schema?.type || ModuleIOValueTypeEnum.string,
+            value: `{{${param.name}}}`
+          });
+        }
+      }
+    }
+    if (item.request) {
+      const properties = item.request?.content?.['application/json']?.schema?.properties;
+      const keys = Object.keys(properties);
+      if (keys.length > 0) {
+        httpNodeBody = JSON.stringify(
+          keys.reduce((acc: any, key) => {
+            acc[key] = `{{${key}}}`;
+            return acc;
+          }, {}),
+          null,
+          2
+        );
+      }
+    }
+    if (customHeader) {
+      const headersObj = (() => {
+        try {
+          return JSON.parse(customHeader) as Record<string, string>;
+        } catch (err) {
+          return {};
+        }
+      })();
+      for (const key in headersObj) {
+        httpNodeHeaders.push({
+          key,
+          type: 'string',
+          // @ts-ignore
+          value: headersObj[key]
+        });
+      }
+    }
+
+    /* Combine complete modules */
+    const modules: ModuleItemType[] = [
+      {
+        moduleId: getNanoid(),
+        name: PluginInputModule.name,
+        intro: PluginInputModule.intro,
+        avatar: PluginInputModule.avatar,
+        flowType: PluginInputModule.flowType,
+        showStatus: PluginInputModule.showStatus,
+        position: {
+          x: 616.4226348688949,
+          y: -165.05298493910115
+        },
+        inputs: [
+          {
+            key: 'pluginStart',
+            type: 'hidden',
+            valueType: 'boolean',
+            label: '插件开始运行',
+            description:
+              '插件开始运行时，会输出一个 True 的标识。有时候，插件不会有额外的的输入，为了顺利的进入下一个阶段，你可以将该值连接到下一个节点的触发器中。',
+            showTargetInApp: true,
+            showTargetInPlugin: true,
+            connected: true
+          },
+          ...pluginInputs
+        ],
+        outputs: [
+          {
+            key: 'pluginStart',
+            label: '插件开始运行',
+            type: 'source',
+            valueType: 'boolean',
+            targets:
+              pluginOutputs.length === 0
+                ? [
+                    {
+                      moduleId: httpId,
+                      key: 'switch'
+                    }
+                  ]
+                : []
+          },
+          ...pluginOutputs
+        ]
+      },
+      {
+        moduleId: pluginOutputId,
+        name: PluginOutputModule.name,
+        intro: PluginOutputModule.intro,
+        avatar: PluginOutputModule.avatar,
+        flowType: PluginOutputModule.flowType,
+        showStatus: PluginOutputModule.showStatus,
+        position: {
+          x: 1607.7142331269126,
+          y: -151.8669210746189
+        },
+        inputs: [
+          {
+            key: pluginOutputKey,
+            valueType: 'string',
+            label: pluginOutputKey,
+            type: 'target',
+            required: true,
+            description: '',
+            edit: true,
+            editField: {
+              key: true,
+              name: true,
+              description: true,
+              required: false,
+              dataType: true,
+              inputType: false
+            },
+            connected: true
+          }
+        ],
+        outputs: [
+          {
+            key: pluginOutputKey,
+            valueType: 'string',
+            label: pluginOutputKey,
+            type: 'source',
+            edit: true,
+            targets: []
+          }
+        ]
+      },
+      {
+        moduleId: httpId,
+        name: HttpModule468.name,
+        intro: HttpModule468.intro,
+        avatar: HttpModule468.avatar,
+        flowType: HttpModule468.flowType,
+        showStatus: true,
+        position: {
+          x: 1042.549746602742,
+          y: -447.77496332641647
+        },
+        inputs: [
+          {
+            key: 'switch',
+            type: 'target',
+            label: 'core.module.input.label.switch',
+            description: 'core.module.input.description.Trigger',
+            valueType: 'any',
+            showTargetInApp: true,
+            showTargetInPlugin: true,
+            connected: false
+          },
+          {
+            key: 'system_httpMethod',
+            type: 'custom',
+            valueType: 'string',
+            label: '',
+            value: item.method.toUpperCase(),
+            required: true,
+            showTargetInApp: false,
+            showTargetInPlugin: false,
+            connected: false
+          },
+          {
+            key: 'system_httpReqUrl',
+            type: 'hidden',
+            valueType: 'string',
+            label: '',
+            description: 'core.module.input.description.Http Request Url',
+            placeholder: 'https://api.ai.com/getInventory',
+            required: false,
+            showTargetInApp: false,
+            showTargetInPlugin: false,
+            value: requestUrl,
+            connected: false
+          },
+          {
+            key: 'system_httpHeader',
+            type: 'custom',
+            valueType: 'any',
+            value: httpNodeHeaders,
+            label: '',
+            description: 'core.module.input.description.Http Request Header',
+            placeholder: 'core.module.input.description.Http Request Header',
+            required: false,
+            showTargetInApp: false,
+            showTargetInPlugin: false,
+            connected: false
+          },
+          {
+            key: 'system_httpParams',
+            type: 'hidden',
+            valueType: 'any',
+            value: httpNodeParams,
+            label: '',
+            required: false,
+            showTargetInApp: false,
+            showTargetInPlugin: false,
+            connected: false
+          },
+          {
+            key: 'system_httpJsonBody',
+            type: 'hidden',
+            valueType: 'any',
+            value: httpNodeBody,
+            label: '',
+            required: false,
+            showTargetInApp: false,
+            showTargetInPlugin: false,
+            connected: false
+          },
+          {
+            key: 'DYNAMIC_INPUT_KEY',
+            type: 'target',
+            valueType: 'any',
+            label: 'core.module.inputType.dynamicTargetInput',
+            description: 'core.module.input.description.dynamic input',
+            required: false,
+            showTargetInApp: false,
+            showTargetInPlugin: true,
+            hideInApp: true,
+            connected: false
+          },
+          {
+            key: 'system_addInputParam',
+            type: 'addInputParam',
+            valueType: 'any',
+            label: '',
+            required: false,
+            showTargetInApp: false,
+            showTargetInPlugin: false,
+            editField: {
+              key: true,
+              description: true,
+              dataType: true
+            },
+            defaultEditField: {
+              label: '',
+              key: '',
+              description: '',
+              inputType: 'target',
+              valueType: 'string'
+            },
+            connected: false
+          },
+          ...httpInputs
+        ],
+        outputs: [
+          {
+            key: 'finish',
+            label: 'core.module.output.label.running done',
+            description: 'core.module.output.description.running done',
+            valueType: 'boolean',
+            type: 'source',
+            targets: []
+          },
+          {
+            key: 'httpRawResponse',
+            label: '原始响应',
+            description: 'HTTP请求的原始响应。只能接受字符串或JSON类型响应数据。',
+            valueType: 'any',
+            type: 'source',
+            targets: [
+              {
+                moduleId: pluginOutputId,
+                key: pluginOutputKey
+              }
+            ]
+          },
+          {
+            key: 'system_addOutputParam',
+            type: 'addOutputParam',
+            valueType: 'any',
+            label: '',
+            targets: [],
+            editField: {
+              key: true,
+              description: true,
+              dataType: true,
+              defaultValue: true
+            },
+            defaultEditField: {
+              label: '',
+              key: '',
+              description: '',
+              outputType: 'source',
+              valueType: 'string'
+            }
+          }
+        ]
+      }
+    ];
+
+    return {
+      name: item.name,
+      avatar: HttpImgUrl,
+      intro: item.description,
+      parentId,
+      type: PluginTypeEnum.http,
+      modules
+    };
+  });
+};
--- a/packages/global/core/plugin/type.d.ts
+++ b/packages/global/core/plugin/type.d.ts
@ -1,6 +1,7 @@
 import { ModuleTemplateTypeEnum } from 'core/module/constants';
 import type { FlowModuleTemplateType, ModuleItemType } from '../module/type.d';
-import { PluginSourceEnum } from './constants';
+import { PluginSourceEnum, PluginTypeEnum } from './constants';
+import { MethodType } from './controller';

 export type PluginItemSchema = {
  _id: string;
@ -12,6 +13,13 @@ export type PluginItemSchema = {
  intro: string;
  updateTime: Date;
  modules: ModuleItemType[];
+  parentId: string;
+  type: `${PluginTypeEnum}`;
+  metadata?: {
+    pluginUid?: string;
+    apiSchemaStr?: string;
+    customHeaders?: string;
+  };
 };

 /* plugin template */
@ -19,7 +27,7 @@ export type PluginTemplateType = PluginRuntimeType & {
  author?: string;
  id: string;
  source: `${PluginSourceEnum}`;
-  templateType: FlowModuleTemplateType['templateType'];
+  templateType: FlowNodeTemplateType['templateType'];
  intro: string;
  modules: ModuleItemType[];
 };
@ -29,5 +37,6 @@ export type PluginRuntimeType = {
  name: string;
  avatar: string;
  showStatus?: boolean;
+  isTool?: boolean;
  modules: ModuleItemType[];
 };
--- a/packages/global/package.json
+++ b/packages/global/package.json
@ -6,11 +6,15 @@
    "dayjs": "^1.11.7",
    "encoding": "^0.1.13",
    "js-tiktoken": "^1.0.7",
+    "openapi-types": "^12.1.3",
    "openai": "4.28.0",
    "nanoid": "^4.0.1",
-    "timezones-list": "^3.0.2"
+    "js-yaml": "^4.1.0",
+    "timezones-list": "^3.0.2",
+    "next": "13.5.2"
  },
  "devDependencies": {
+    "@types/js-yaml": "^4.0.9",
    "@types/node": "^20.8.5"
  }
 }
--- a/projects/app/src/service/common/censor/index.ts
+++ b/projects/app/src/service/common/censor/index.ts
@ -1,4 +1,4 @@
-import { POST } from '@fastgpt/service/common/api/plusRequest';
+import { POST } from './plusRequest';

 export const postTextCensor = (data: { text: string }) =>
  POST<{ code?: number; message: string }>('/common/censor/text_baidu', data)
--- a/packages/service/common/api/serverRequest.ts
+++ b/packages/service/common/api/serverRequest.ts
@ -0,0 +1,114 @@
+import { SERVICE_LOCAL_HOST } from '../system/tools';
+import axios, { Method, InternalAxiosRequestConfig, AxiosResponse } from 'axios';
+
+interface ConfigType {
+  headers?: { [key: string]: string };
+  hold?: boolean;
+  timeout?: number;
+}
+interface ResponseDataType {
+  code: number;
+  message: string;
+  data: any;
+}
+
+/**
+ * 请求开始
+ */
+function requestStart(config: InternalAxiosRequestConfig): InternalAxiosRequestConfig {
+  return config;
+}
+
+/**
+ * 请求成功,检查请求头
+ */
+function responseSuccess(response: AxiosResponse<ResponseDataType>) {
+  return response;
+}
+/**
+ * 响应数据检查
+ */
+function checkRes(data: ResponseDataType) {
+  if (data === undefined) {
+    console.log('error->', data, 'data is empty');
+    return Promise.reject('服务器异常');
+  } else if (data?.code && (data.code < 200 || data.code >= 400)) {
+    return Promise.reject(data);
+  }
+  return data.data;
+}
+
+/**
+ * 响应错误
+ */
+function responseError(err: any) {
+  if (!err) {
+    return Promise.reject({ message: '未知错误' });
+  }
+  if (typeof err === 'string') {
+    return Promise.reject({ message: err });
+  }
+
+  if (err?.response?.data) {
+    return Promise.reject(err?.response?.data);
+  }
+  return Promise.reject(err);
+}
+
+/* 创建请求实例 */
+const instance = axios.create({
+  timeout: 60000, // 超时时间
+  headers: {
+    'content-type': 'application/json',
+    'Cache-Control': 'no-cache'
+  }
+});
+
+/* 请求拦截 */
+instance.interceptors.request.use(requestStart, (err) => Promise.reject(err));
+/* 响应拦截 */
+instance.interceptors.response.use(responseSuccess, (err) => Promise.reject(err));
+
+export function request(url: string, data: any, config: ConfigType, method: Method): any {
+  /* 去空 */
+  for (const key in data) {
+    if (data[key] === null || data[key] === undefined) {
+      delete data[key];
+    }
+  }
+
+  return instance
+    .request({
+      baseURL: `http://${SERVICE_LOCAL_HOST}`,
+      url,
+      method,
+      data: ['POST', 'PUT'].includes(method) ? data : null,
+      params: !['POST', 'PUT'].includes(method) ? data : null,
+      ...config // custom config
+    })
+    .then((res) => checkRes(res.data))
+    .catch((err) => responseError(err));
+}
+
+/**
+ * api请求方式
+ * @param {String} url
+ * @param {Any} params
+ * @param {Object} config
+ * @returns
+ */
+export function GET<T = undefined>(url: string, params = {}, config: ConfigType = {}): Promise<T> {
+  return request(url, params, config, 'GET');
+}
+
+export function POST<T = undefined>(url: string, data = {}, config: ConfigType = {}): Promise<T> {
+  return request(url, data, config, 'POST');
+}
+
+export function PUT<T = undefined>(url: string, data = {}, config: ConfigType = {}): Promise<T> {
+  return request(url, data, config, 'PUT');
+}
+
+export function DELETE<T = undefined>(url: string, data = {}, config: ConfigType = {}): Promise<T> {
+  return request(url, data, config, 'DELETE');
+}
--- a/packages/service/common/string/jieba.ts
+++ b/packages/service/common/string/jieba.ts
@ -1,4 +1,6 @@
-export const stopWords = new Set([
+import { cut } from '@node-rs/jieba';
+
+const stopWords = new Set([
  '--',
  '?',
  '“',
@ -1506,3 +1508,14 @@ export const stopWords = new Set([
    'i'
  ]
 ]);
+
+export function jiebaSplit({ text }: { text: string }) {
+  const tokens = cut(text, true);
+
+  return (
+    tokens
+      .map((item) => item.replace(/[\u3000-\u303f\uff00-\uffef]/g, '').trim())
+      .filter((item) => item && !stopWords.has(item))
+      .join(' ') || ''
+  );
+}
--- a/packages/service/common/vectorStore/pg/controller.ts
+++ b/packages/service/common/vectorStore/pg/controller.ts
@ -139,11 +139,11 @@ export const embeddingRecall = async (
    const results: any = await PgClient.query(
      `BEGIN;
        SET LOCAL hnsw.ef_search = ${efSearch};
-        select id, collection_id, (vector <#> '[${vectors[0]}]') * -1 AS score 
+        select id, collection_id, vector <#> '[${vectors[0]}]' AS score 
          from ${PgDatasetTableName} 
          where dataset_id IN (${datasetIds.map((id) => `'${String(id)}'`).join(',')})
              AND vector <#> '[${vectors[0]}]' < -${similarity}
-          order by score desc limit ${limit};
+          order by score limit ${limit};
        COMMIT;`
    );

@ -153,7 +153,7 @@ export const embeddingRecall = async (
      results: rows.map((item) => ({
        id: item.id,
        collectionId: item.collection_id,
-        score: item.score
+        score: item.score * -1
      }))
    };
  } catch (error) {
--- a/packages/service/core/ai/functions/createQuestionGuide.ts
+++ b/packages/service/core/ai/functions/createQuestionGuide.ts
@ -2,7 +2,7 @@ import type { ChatCompletionMessageParam } from '@fastgpt/global/core/ai/type.d'
 import { getAIApi } from '../config';
 import { countGptMessagesTokens } from '@fastgpt/global/common/string/tiktoken';

-export const Prompt_QuestionGuide = `我不太清楚问你什么问题，请帮我生成 3 个问题，引导我继续提问。问题的长度应小于20个字符，按 JSON 格式返回: ["问题1", "问题2", "问题3"]`;
+export const Prompt_QuestionGuide = `你是一个AI智能助手，可以回答和解决我的问题。请结合前面的对话记录，帮我生成 3 个问题，引导我继续提问。问题的长度应小于20个字符，按 JSON 格式返回: ["问题1", "问题2", "问题3"]`;

 export async function createQuestionGuide({
  messages,
--- a/packages/service/core/ai/functions/queryExtension.ts
+++ b/packages/service/core/ai/functions/queryExtension.ts
@ -9,7 +9,7 @@ import { ChatCompletionMessageParam } from '@fastgpt/global/core/ai/type';
    可以根据上下文，消除指代性问题以及扩展问题，利于检索。
 */

-const defaultPrompt = `作为一个向量检索助手，你的任务是结合历史记录，从不同角度，为“原问题”生成个不同版本的“检索词”，从而提高向量检索的语义丰富度，提高向量检索的精度。生成的问题要求指向对象清晰明确，并与原问题语言相同。例如：
+const defaultPrompt = `作为一个向量检索助手，你的任务是结合历史记录，从不同角度，为“原问题”生成个不同版本的“检索词”，从而提高向量检索的语义丰富度，提高向量检索的精度。生成的问题要求指向对象清晰明确，并与“原问题语言相同”。例如：
 历史记录: 
 """
 """
--- a/projects/app/src/service/core/ai/rerank.ts
+++ b/projects/app/src/service/core/ai/rerank.ts
@ -1,5 +1,5 @@
-import { PostReRankProps, PostReRankResponse } from '@fastgpt/global/core/ai/api';
-import { POST } from '@/service/common/api/request';
+import { PostReRankProps, PostReRankResponse } from '@fastgpt/global/core/ai/api.d';
+import { POST } from '../../../common/api/serverRequest';

 export function reRankRecall({ query, inputs }: PostReRankProps) {
  const model = global.reRankModels[0];
--- a/packages/service/core/app/schema.ts
+++ b/packages/service/core/app/schema.ts
@ -34,10 +34,6 @@ const AppSchema = new Schema({
    default: 'advanced',
    enum: Object.keys(AppTypeMap)
  },
-  simpleTemplateId: {
-    type: String,
-    required: true
-  },
  avatar: {
    type: String,
    default: '/icon/logo.svg'
--- a/packages/service/core/chat/utils.ts
+++ b/packages/service/core/chat/utils.ts
@ -8,6 +8,15 @@ import axios from 'axios';
 import { ChatCompletionRequestMessageRoleEnum } from '@fastgpt/global/core/ai/constants';

 /* slice chat context by tokens */
+const filterEmptyMessages = (messages: ChatCompletionMessageParam[]) => {
+  return messages.filter((item) => {
+    if (item.role === ChatCompletionRequestMessageRoleEnum.System) return !!item.content;
+    if (item.role === ChatCompletionRequestMessageRoleEnum.User) return !!item.content;
+    if (item.role === ChatCompletionRequestMessageRoleEnum.Assistant)
+      return !!item.content || !!item.function_call || !!item.tool_calls;
+    return true;
+  });
+};
 export function filterGPTMessageByMaxTokens({
  messages = [],
  maxTokens
@ -38,7 +47,7 @@ export function filterGPTMessageByMaxTokens({

  // If the text length is less than half of the maximum token, no calculation is required
  if (rawTextLen < maxTokens * 0.5) {
-    return messages;
+    return filterEmptyMessages(messages);
  }

  // filter startWith system prompt
@ -81,7 +90,7 @@ export function filterGPTMessageByMaxTokens({
    }
  }

-  return [...systemPrompts, ...chats];
+  return filterEmptyMessages([...systemPrompts, ...chats]);
 }
 export const formatGPTMessagesInRequestBefore = (messages: ChatCompletionMessageParam[]) => {
  return messages
--- a/packages/service/core/dataset/schema.ts
+++ b/packages/service/core/dataset/schema.ts
@ -67,7 +67,7 @@ const DatasetSchema = new Schema({
  agentModel: {
    type: String,
    required: true,
-    default: 'gpt-3.5-turbo-16k'
+    default: 'gpt-3.5-turbo'
  },
  intro: {
    type: String,
--- a/packages/service/core/dataset/search/controller.ts
+++ b/packages/service/core/dataset/search/controller.ts
@ -0,0 +1,407 @@
+import {
+  DatasetSearchModeEnum,
+  DatasetSearchModeMap,
+  SearchScoreTypeEnum
+} from '@fastgpt/global/core/dataset/constants';
+import { recallFromVectorStore } from '../../../common/vectorStore/controller';
+import { getVectorsByText } from '../../ai/embedding';
+import { getVectorModel } from '../../ai/model';
+import { MongoDatasetData } from '../data/schema';
+import {
+  DatasetDataSchemaType,
+  DatasetDataWithCollectionType,
+  SearchDataResponseItemType
+} from '@fastgpt/global/core/dataset/type';
+import { MongoDatasetCollection } from '../collection/schema';
+import { reRankRecall } from '../../../core/ai/rerank';
+import { countPromptTokens } from '@fastgpt/global/common/string/tiktoken';
+import { datasetSearchResultConcat } from '@fastgpt/global/core/dataset/search/utils';
+import { hashStr } from '@fastgpt/global/common/string/tools';
+import { jiebaSplit } from '../../../common/string/jieba';
+
+type SearchDatasetDataProps = {
+  teamId: string;
+  model: string;
+  similarity?: number; // min distance
+  limit: number; // max Token limit
+  datasetIds: string[];
+  searchMode?: `${DatasetSearchModeEnum}`;
+  usingReRank?: boolean;
+  reRankQuery: string;
+  queries: string[];
+};
+
+export async function searchDatasetData(props: SearchDatasetDataProps) {
+  let {
+    teamId,
+    reRankQuery,
+    queries,
+    model,
+    similarity = 0,
+    limit: maxTokens,
+    searchMode = DatasetSearchModeEnum.embedding,
+    usingReRank = false,
+    datasetIds = []
+  } = props;
+
+  /* init params */
+  searchMode = DatasetSearchModeMap[searchMode] ? searchMode : DatasetSearchModeEnum.embedding;
+  usingReRank = usingReRank && global.reRankModels.length > 0;
+
+  // Compatible with topk limit
+  if (maxTokens < 50) {
+    maxTokens = 1500;
+  }
+  let set = new Set<string>();
+  let usingSimilarityFilter = false;
+
+  /* function */
+  const countRecallLimit = () => {
+    if (searchMode === DatasetSearchModeEnum.embedding) {
+      return {
+        embeddingLimit: 150,
+        fullTextLimit: 0
+      };
+    }
+    if (searchMode === DatasetSearchModeEnum.fullTextRecall) {
+      return {
+        embeddingLimit: 0,
+        fullTextLimit: 150
+      };
+    }
+    return {
+      embeddingLimit: 100,
+      fullTextLimit: 80
+    };
+  };
+  const embeddingRecall = async ({ query, limit }: { query: string; limit: number }) => {
+    const { vectors, tokens } = await getVectorsByText({
+      model: getVectorModel(model),
+      input: query
+    });
+
+    const { results } = await recallFromVectorStore({
+      vectors,
+      limit,
+      datasetIds,
+      efSearch: global.systemEnv?.pgHNSWEfSearch
+    });
+
+    // get q and a
+    const dataList = (await MongoDatasetData.find(
+      {
+        teamId,
+        datasetId: { $in: datasetIds },
+        'indexes.dataId': { $in: results.map((item) => item.id?.trim()) }
+      },
+      'datasetId collectionId q a chunkIndex indexes'
+    )
+      .populate('collectionId', 'name fileId rawLink')
+      .lean()) as DatasetDataWithCollectionType[];
+
+    // add score to data(It's already sorted. The first one is the one with the most points)
+    const concatResults = dataList.map((data) => {
+      const dataIdList = data.indexes.map((item) => item.dataId);
+
+      const maxScoreResult = results.find((item) => {
+        return dataIdList.includes(item.id);
+      });
+
+      return {
+        ...data,
+        score: maxScoreResult?.score || 0
+      };
+    });
+
+    concatResults.sort((a, b) => b.score - a.score);
+
+    const formatResult = concatResults
+      .map((data, index) => {
+        if (!data.collectionId) {
+          console.log('Collection is not found', data);
+        }
+
+        const result: SearchDataResponseItemType = {
+          id: String(data._id),
+          q: data.q,
+          a: data.a,
+          chunkIndex: data.chunkIndex,
+          datasetId: String(data.datasetId),
+          collectionId: String(data.collectionId?._id),
+          sourceName: data.collectionId?.name || '',
+          sourceId: data.collectionId?.fileId || data.collectionId?.rawLink,
+          score: [{ type: SearchScoreTypeEnum.embedding, value: data.score, index }]
+        };
+
+        return result;
+      })
+      .filter((item) => item !== null) as SearchDataResponseItemType[];
+
+    return {
+      embeddingRecallResults: formatResult,
+      tokens
+    };
+  };
+  const fullTextRecall = async ({
+    query,
+    limit
+  }: {
+    query: string;
+    limit: number;
+  }): Promise<{
+    fullTextRecallResults: SearchDataResponseItemType[];
+    tokenLen: number;
+  }> => {
+    if (limit === 0) {
+      return {
+        fullTextRecallResults: [],
+        tokenLen: 0
+      };
+    }
+
+    let searchResults = (
+      await Promise.all(
+        datasetIds.map((id) =>
+          MongoDatasetData.find(
+            {
+              teamId,
+              datasetId: id,
+              $text: { $search: jiebaSplit({ text: query }) }
+            },
+            {
+              score: { $meta: 'textScore' },
+              _id: 1,
+              datasetId: 1,
+              collectionId: 1,
+              q: 1,
+              a: 1,
+              chunkIndex: 1
+            }
+          )
+            .sort({ score: { $meta: 'textScore' } })
+            .limit(limit)
+            .lean()
+        )
+      )
+    ).flat() as (DatasetDataSchemaType & { score: number })[];
+
+    // resort
+    searchResults.sort((a, b) => b.score - a.score);
+    searchResults.slice(0, limit);
+
+    const collections = await MongoDatasetCollection.find(
+      {
+        _id: { $in: searchResults.map((item) => item.collectionId) }
+      },
+      '_id name fileId rawLink'
+    );
+
+    return {
+      fullTextRecallResults: searchResults.map((item, index) => {
+        const collection = collections.find((col) => String(col._id) === String(item.collectionId));
+        return {
+          id: String(item._id),
+          datasetId: String(item.datasetId),
+          collectionId: String(item.collectionId),
+          sourceName: collection?.name || '',
+          sourceId: collection?.fileId || collection?.rawLink,
+          q: item.q,
+          a: item.a,
+          chunkIndex: item.chunkIndex,
+          indexes: item.indexes,
+          score: [{ type: SearchScoreTypeEnum.fullText, value: item.score, index }]
+        };
+      }),
+      tokenLen: 0
+    };
+  };
+  const reRankSearchResult = async ({
+    data,
+    query
+  }: {
+    data: SearchDataResponseItemType[];
+    query: string;
+  }): Promise<SearchDataResponseItemType[]> => {
+    try {
+      const results = await reRankRecall({
+        query,
+        inputs: data.map((item) => ({
+          id: item.id,
+          text: `${item.q}\n${item.a}`
+        }))
+      });
+
+      if (results.length === 0) {
+        usingReRank = false;
+        return [];
+      }
+
+      // add new score to data
+      const mergeResult = results
+        .map((item, index) => {
+          const target = data.find((dataItem) => dataItem.id === item.id);
+          if (!target) return null;
+          const score = item.score || 0;
+
+          return {
+            ...target,
+            score: [{ type: SearchScoreTypeEnum.reRank, value: score, index }]
+          };
+        })
+        .filter(Boolean) as SearchDataResponseItemType[];
+
+      return mergeResult;
+    } catch (error) {
+      usingReRank = false;
+      return [];
+    }
+  };
+  const filterResultsByMaxTokens = (list: SearchDataResponseItemType[], maxTokens: number) => {
+    const results: SearchDataResponseItemType[] = [];
+    let totalTokens = 0;
+
+    for (let i = 0; i < list.length; i++) {
+      const item = list[i];
+      totalTokens += countPromptTokens(item.q + item.a);
+      if (totalTokens > maxTokens + 500) {
+        break;
+      }
+      results.push(item);
+      if (totalTokens > maxTokens) {
+        break;
+      }
+    }
+
+    return results.length === 0 ? list.slice(0, 1) : results;
+  };
+  const multiQueryRecall = async ({
+    embeddingLimit,
+    fullTextLimit
+  }: {
+    embeddingLimit: number;
+    fullTextLimit: number;
+  }) => {
+    // multi query recall
+    const embeddingRecallResList: SearchDataResponseItemType[][] = [];
+    const fullTextRecallResList: SearchDataResponseItemType[][] = [];
+    let totalTokens = 0;
+
+    await Promise.all(
+      queries.map(async (query) => {
+        const [{ tokens, embeddingRecallResults }, { fullTextRecallResults }] = await Promise.all([
+          embeddingRecall({
+            query,
+            limit: embeddingLimit
+          }),
+          fullTextRecall({
+            query,
+            limit: fullTextLimit
+          })
+        ]);
+        totalTokens += tokens;
+
+        embeddingRecallResList.push(embeddingRecallResults);
+        fullTextRecallResList.push(fullTextRecallResults);
+      })
+    );
+
+    // rrf concat
+    const rrfEmbRecall = datasetSearchResultConcat(
+      embeddingRecallResList.map((list) => ({ k: 60, list }))
+    ).slice(0, embeddingLimit);
+    const rrfFTRecall = datasetSearchResultConcat(
+      fullTextRecallResList.map((list) => ({ k: 60, list }))
+    ).slice(0, fullTextLimit);
+
+    return {
+      tokens: totalTokens,
+      embeddingRecallResults: rrfEmbRecall,
+      fullTextRecallResults: rrfFTRecall
+    };
+  };
+
+  /* main step */
+  // count limit
+  const { embeddingLimit, fullTextLimit } = countRecallLimit();
+
+  // recall
+  const { embeddingRecallResults, fullTextRecallResults, tokens } = await multiQueryRecall({
+    embeddingLimit,
+    fullTextLimit
+  });
+
+  // ReRank results
+  const reRankResults = await (async () => {
+    if (!usingReRank) return [];
+
+    set = new Set<string>(embeddingRecallResults.map((item) => item.id));
+    const concatRecallResults = embeddingRecallResults.concat(
+      fullTextRecallResults.filter((item) => !set.has(item.id))
+    );
+
+    // remove same q and a data
+    set = new Set<string>();
+    const filterSameDataResults = concatRecallResults.filter((item) => {
+      // 删除所有的标点符号与空格等，只对文本进行比较
+      const str = hashStr(`${item.q}${item.a}`.replace(/[^\p{L}\p{N}]/gu, ''));
+      if (set.has(str)) return false;
+      set.add(str);
+      return true;
+    });
+    return reRankSearchResult({
+      query: reRankQuery,
+      data: filterSameDataResults
+    });
+  })();
+
+  // embedding recall and fullText recall rrf concat
+  const rrfConcatResults = datasetSearchResultConcat([
+    { k: 60, list: embeddingRecallResults },
+    { k: 60, list: fullTextRecallResults },
+    { k: 58, list: reRankResults }
+  ]);
+
+  // remove same q and a data
+  set = new Set<string>();
+  const filterSameDataResults = rrfConcatResults.filter((item) => {
+    // 删除所有的标点符号与空格等，只对文本进行比较
+    const str = hashStr(`${item.q}${item.a}`.replace(/[^\p{L}\p{N}]/gu, ''));
+    if (set.has(str)) return false;
+    set.add(str);
+    return true;
+  });
+
+  // score filter
+  const scoreFilter = (() => {
+    if (usingReRank) {
+      usingSimilarityFilter = true;
+
+      return filterSameDataResults.filter((item) => {
+        const reRankScore = item.score.find((item) => item.type === SearchScoreTypeEnum.reRank);
+        if (reRankScore && reRankScore.value < similarity) return false;
+        return true;
+      });
+    }
+    if (searchMode === DatasetSearchModeEnum.embedding) {
+      usingSimilarityFilter = true;
+      return filterSameDataResults.filter((item) => {
+        const embeddingScore = item.score.find(
+          (item) => item.type === SearchScoreTypeEnum.embedding
+        );
+        if (embeddingScore && embeddingScore.value < similarity) return false;
+        return true;
+      });
+    }
+    return filterSameDataResults;
+  })();
+
+  return {
+    searchRes: filterResultsByMaxTokens(scoreFilter, maxTokens),
+    tokens,
+    searchMode,
+    limit: maxTokens,
+    similarity,
+    usingReRank,
+    usingSimilarityFilter
+  };
+}
--- a/packages/service/core/dataset/training/controller.ts
+++ b/packages/service/core/dataset/training/controller.ts
@ -31,16 +31,14 @@ export async function pushDataListToTrainingQueue({
  data,
  prompt,
  billId,
-  trainingMode = TrainingModeEnum.chunk,
-
-  vectorModelList = [],
-  datasetModelList = []
+  trainingMode = TrainingModeEnum.chunk
 }: {
  teamId: string;
  tmbId: string;
-  vectorModelList: VectorModelItemType[];
-  datasetModelList: LLMModelItemType[];
 } & PushDatasetDataProps): Promise<PushDatasetDataResponse> {
+  const vectorModelList = global.vectorModels;
+  const datasetModelList = global.llmModels;
+
  const {
    datasetId: { _id: datasetId, vectorModel, agentModel }
  } = await getCollectionWithDataset(collectionId);
@ -48,11 +46,11 @@ export async function pushDataListToTrainingQueue({
  const checkModelValid = async () => {
    const agentModelData = datasetModelList?.find((item) => item.model === agentModel);
    if (!agentModelData) {
-      return Promise.reject(`Vector model ${agentModel} is inValid`);
+      return Promise.reject(`File model ${agentModel} is inValid`);
    }
    const vectorModelData = vectorModelList?.find((item) => item.model === vectorModel);
    if (!vectorModelData) {
-      return Promise.reject(`File model ${vectorModel} is inValid`);
+      return Promise.reject(`Vector model ${vectorModel} is inValid`);
    }

    if (trainingMode === TrainingModeEnum.chunk) {
--- a/packages/service/core/dataset/training/schema.ts
+++ b/packages/service/core/dataset/training/schema.ts
@ -13,11 +13,6 @@ import {
 export const DatasetTrainingCollectionName = 'dataset.trainings';

 const TrainingDataSchema = new Schema({
-  userId: {
-    // abandon
-    type: Schema.Types.ObjectId,
-    ref: 'user'
-  },
  teamId: {
    type: Schema.Types.ObjectId,
    ref: TeamCollectionName,
@ -100,7 +95,7 @@ try {
  // lock training data; delete training data
  TrainingDataSchema.index({ teamId: 1, collectionId: 1 });
  // get training data and sort
-  TrainingDataSchema.index({ lockTime: 1, mode: 1, weight: -1 });
+  TrainingDataSchema.index({ mode: 1, lockTime: 1, weight: -1 });
  TrainingDataSchema.index({ expireAt: 1 }, { expireAfterSeconds: 7 * 24 * 60 * 60 }); // 7 days
 } catch (error) {
  console.log(error);
--- a/packages/service/core/plugin/controller.ts
+++ b/packages/service/core/plugin/controller.ts
@ -1,10 +1,11 @@
 import { MongoPlugin } from './schema';
-import { FlowModuleTemplateType } from '@fastgpt/global/core/module/type';
+import { FlowNodeTemplateType } from '@fastgpt/global/core/module/type';
 import { FlowNodeTypeEnum } from '@fastgpt/global/core/module/node/constant';
 import { plugin2ModuleIO } from '@fastgpt/global/core/module/utils';
 import { PluginSourceEnum } from '@fastgpt/global/core/plugin/constants';
 import type { PluginRuntimeType, PluginTemplateType } from '@fastgpt/global/core/plugin/type.d';
-import { ModuleTemplateTypeEnum } from '@fastgpt/global/core/module/constants';
+import { FlowNodeTemplateTypeEnum } from '@fastgpt/global/core/module/constants';
+import type { PluginItemSchema } from '@fastgpt/global/core/plugin/type.d';

 /* 
  plugin id rule:
@ -48,7 +49,7 @@ const getPluginTemplateById = async (id: string): Promise<PluginTemplateType> =>
      showStatus: true,
      source: PluginSourceEnum.personal,
      modules: item.modules,
-      templateType: ModuleTemplateTypeEnum.personalPlugin
+      templateType: FlowNodeTemplateTypeEnum.personalPlugin
    };
  }
  return Promise.reject('plugin not found');
@ -59,7 +60,7 @@ export async function getPluginPreviewModule({
  id
 }: {
  id: string;
-}): Promise<FlowModuleTemplateType> {
+}): Promise<FlowNodeTemplateType> {
  const plugin = await getPluginTemplateById(id);

  return {
@ -70,6 +71,7 @@ export async function getPluginPreviewModule({
    name: plugin.name,
    intro: plugin.intro,
    showStatus: plugin.showStatus,
+    isTool: plugin.isTool,
    ...plugin2ModuleIO(plugin.id, plugin.modules)
  };
 }
--- a/packages/service/core/plugin/schema.ts
+++ b/packages/service/core/plugin/schema.ts
@ -1,3 +1,4 @@
+import { pluginTypeMap } from '@fastgpt/global/core/plugin/constants';
 import { connectionMongo, type Model } from '../../common/mongo';
 const { Schema, model, models } = connectionMongo;
 import type { PluginItemSchema } from '@fastgpt/global/core/plugin/type.d';
@ -9,9 +10,10 @@ import {
 export const PluginCollectionName = 'plugins';

 const PluginSchema = new Schema({
-  userId: {
+  parentId: {
    type: Schema.Types.ObjectId,
-    ref: 'user'
+    ref: PluginCollectionName,
+    default: null
  },
  teamId: {
    type: Schema.Types.ObjectId,
@ -23,6 +25,11 @@ const PluginSchema = new Schema({
    ref: TeamMemberCollectionName,
    required: true
  },
+  type: {
+    type: String,
+    enum: Object.keys(pluginTypeMap),
+    required: true
+  },
  name: {
    type: String,
    required: true
@ -42,11 +49,19 @@ const PluginSchema = new Schema({
  modules: {
    type: Array,
    default: []
+  },
+  metadata: {
+    type: {
+      pluginUid: String,
+      apiSchemaStr: String,
+      customHeaders: String
+    }
  }
 });

 try {
-  PluginSchema.index({ tmbId: 1 });
+  PluginSchema.index({ teamId: 1, parentId: 1 });
+  PluginSchema.index({ teamId: 1, name: 1, intro: 1 });
 } catch (error) {
  console.log(error);
 }
--- a/packages/service/core/workflow/dispatch/agent/classifyQuestion.ts
+++ b/packages/service/core/workflow/dispatch/agent/classifyQuestion.ts
@ -1,22 +1,22 @@
 import { chats2GPTMessages } from '@fastgpt/global/core/chat/adapt';
-import { filterGPTMessageByMaxTokens } from '@fastgpt/service/core/chat/utils';
+import { filterGPTMessageByMaxTokens } from '../../../chat/utils';
 import {
  countGptMessagesTokens,
  countMessagesTokens
 } from '@fastgpt/global/common/string/tiktoken';
 import type { ChatItemType } from '@fastgpt/global/core/chat/type.d';
 import { ChatItemValueTypeEnum, ChatRoleEnum } from '@fastgpt/global/core/chat/constants';
-import { getAIApi } from '@fastgpt/service/core/ai/config';
+import { getAIApi } from '../../../ai/config';
 import type { ClassifyQuestionAgentItemType } from '@fastgpt/global/core/module/type.d';
 import { ModuleInputKeyEnum } from '@fastgpt/global/core/module/constants';
 import { DispatchNodeResponseKeyEnum } from '@fastgpt/global/core/module/runtime/constants';
 import type { ModuleDispatchProps } from '@fastgpt/global/core/module/type.d';
 import { replaceVariable } from '@fastgpt/global/common/string/tools';
-import { Prompt_CQJson } from '@/global/core/prompt/agent';
+import { Prompt_CQJson } from '@fastgpt/global/core/ai/prompt/agent';
 import { LLMModelItemType } from '@fastgpt/global/core/ai/model.d';
-import { ModelTypeEnum, getLLMModel } from '@fastgpt/service/core/ai/model';
+import { ModelTypeEnum, getLLMModel } from '../../../ai/model';
 import { getHistories } from '../utils';
-import { formatModelChars2Points } from '@fastgpt/service/support/wallet/usage/utils';
+import { formatModelChars2Points } from '../../../../support/wallet/usage/utils';
 import { ChatCompletionRequestMessageRoleEnum } from '@fastgpt/global/core/ai/constants';
 import {
  ChatCompletionCreateParams,
--- a/packages/service/core/workflow/dispatch/agent/extract.ts
+++ b/packages/service/core/workflow/dispatch/agent/extract.ts
@ -1,22 +1,22 @@
 import { chats2GPTMessages } from '@fastgpt/global/core/chat/adapt';
-import { filterGPTMessageByMaxTokens } from '@fastgpt/service/core/chat/utils';
+import { filterGPTMessageByMaxTokens } from '../../../chat/utils';
 import type { ChatItemType } from '@fastgpt/global/core/chat/type.d';
 import {
  countGptMessagesTokens,
  countMessagesTokens
 } from '@fastgpt/global/common/string/tiktoken';
 import { ChatItemValueTypeEnum, ChatRoleEnum } from '@fastgpt/global/core/chat/constants';
-import { getAIApi } from '@fastgpt/service/core/ai/config';
+import { getAIApi } from '../../../ai/config';
 import type { ContextExtractAgentItemType } from '@fastgpt/global/core/module/type';
 import { ModuleInputKeyEnum, ModuleOutputKeyEnum } from '@fastgpt/global/core/module/constants';
 import { DispatchNodeResponseKeyEnum } from '@fastgpt/global/core/module/runtime/constants';
 import type { ModuleDispatchProps } from '@fastgpt/global/core/module/type.d';
-import { Prompt_ExtractJson } from '@/global/core/prompt/agent';
+import { Prompt_ExtractJson } from '@fastgpt/global/core/ai/prompt/agent';
 import { replaceVariable } from '@fastgpt/global/common/string/tools';
 import { LLMModelItemType } from '@fastgpt/global/core/ai/model.d';
 import { getHistories } from '../utils';
-import { ModelTypeEnum, getLLMModel } from '@fastgpt/service/core/ai/model';
-import { formatModelChars2Points } from '@fastgpt/service/support/wallet/usage/utils';
+import { ModelTypeEnum, getLLMModel } from '../../../ai/model';
+import { formatModelChars2Points } from '../../../../support/wallet/usage/utils';
 import json5 from 'json5';
 import {
  ChatCompletionCreateParams,
@ -42,7 +42,7 @@ type Response = DispatchNodeResultType<{

 type ActionProps = Props & { extractModel: LLMModelItemType };

-const agentFunName = 'extract_json_data';
+const agentFunName = 'request_function';

 export async function dispatchContentExtract(props: Props): Promise<Response> {
  const {
@ -156,13 +156,15 @@ const getFunctionCallSchema = ({
        {
          type: ChatItemValueTypeEnum.text,
          text: {
-            content: `你的任务是根据上下文获取适当的 JSON 字符串。要求：
-          """
-          - 字符串不要换行。
-          - 结合上下文和当前问题进行获取。
-          """
+            content: `我正在执行一个函数，需要你提供一些参数，请以 JSON 字符串格式返回这些参数，要求：
+"""
+${description ? `- ${description}` : ''}
+- 不是每个参数都是必须生成的，如果没有合适的参数值，不要生成该参数，或返回空字符串。
+- 需要结合前面的对话内容，一起生成合适的参数。
+"""

-          当前问题: "${content}"`
+本次输入内容: ${content}
+            `
          }
        }
      ]
@ -191,7 +193,7 @@ const getFunctionCallSchema = ({
  // function body
  const agentFunction = {
    name: agentFunName,
-    description,
+    description: '需要执行的函数',
    parameters: {
      type: 'object',
      properties
@ -249,7 +251,6 @@ const toolChoice = async (props: ActionProps) => {
      tool_calls: response.choices?.[0]?.message?.tool_calls
    }
  ];
-
  return {
    tokens: countGptMessagesTokens(completeMessages, tools),
    arg
@ -351,18 +352,24 @@ Human: ${content}`
  const start = answer.indexOf('{');
  const end = answer.lastIndexOf('}');

-  if (start === -1 || end === -1)
+  if (start === -1 || end === -1) {
    return {
      rawResponse: answer,
      tokens: countMessagesTokens(messages),
      arg: {}
    };
+  }
+
+  const jsonStr = answer
+    .substring(start, end + 1)
+    .replace(/(\\n|\\)/g, '')
+    .replace(/  /g, '');

  try {
    return {
      rawResponse: answer,
      tokens: countMessagesTokens(messages),
-      arg: json5.parse(answer) as Record<string, any>
+      arg: json5.parse(jsonStr) as Record<string, any>
    };
  } catch (error) {
    console.log(error);
--- a/packages/service/core/workflow/dispatch/agent/runTool/constants.ts
+++ b/packages/service/core/workflow/dispatch/agent/runTool/constants.ts
@ -0,0 +1,36 @@
+export const Prompt_Tool_Call = `<Instruction>
+你是一个智能机器人，除了可以回答用户问题外，你还掌握工具的使用能力。有时候，你可以依赖工具的运行结果，来更准确的回答用户。
+下面是你可以使用的工具，使用 JSON Schema 的格式声明，其中 toolId 是工具的 description 是工具的描述，parameters 是工具的参数，包括参数的类型和描述，required 是必填参数的列表。
+
+"""
+{{toolsPrompt}}
+"""
+
+接下来，请你根据工具描述，决定回答问题或是使用工具。在完成任务过程中，USER代表用户的输入，TOOL_RESPONSE代表工具运行结果。ASSISTANT 代表你的输出。
+你的每次输出都必须以0,1开头，代表是否需要调用工具：
+0: 不使用工具，直接回答内容。
+1: 使用工具，返回工具调用的参数。
+
+例如：
+
+USER: 你好呀
+ANSWER: 0: 你好，有什么可以帮助你的么？
+USER: 今天杭州的天气如何
+ANSWER: 1: {"toolId":"w2121",arguments:{"city": "杭州"}}
+TOOL_RESPONSE: """
+晴天......
+"""
+ANSWER: 0: 今天杭州是晴天。
+USER: 今天杭州的天气适合去哪里玩？
+ANSWER: 1: {"toolId":"as21da",arguments:{"query": "杭州 天气 去哪里玩"}}
+TOOL_RESPONSE: """
+晴天. 西湖、灵隐寺、千岛湖……
+"""
+ANSWER: 0: 今天杭州是晴天，适合去西湖、灵隐寺、千岛湖等地玩。
+</Instruction>
+
+现在，我们开始吧！
+
+USER: {{question}}
+ANSWER: 
+`;
--- a/packages/service/core/workflow/dispatch/agent/runTool/functionCall.ts
+++ b/packages/service/core/workflow/dispatch/agent/runTool/functionCall.ts
@ -1,6 +1,6 @@
 import { LLMModelItemType } from '@fastgpt/global/core/ai/model.d';
-import { getAIApi } from '@fastgpt/service/core/ai/config';
-import { filterGPTMessageByMaxTokens } from '@fastgpt/service/core/chat/utils';
+import { getAIApi } from '../../../../ai/config';
+import { filterGPTMessageByMaxTokens } from '../../../../chat/utils';
 import {
  ChatCompletion,
  StreamChatType,
@ -15,7 +15,7 @@ import {
  responseWrite,
  responseWriteController,
  responseWriteNodeStatus
-} from '@fastgpt/service/common/response';
+} from '../../../../../common/response';
 import { SseResponseEventEnum } from '@fastgpt/global/core/module/runtime/constants';
 import { textAdaptGptResponse } from '@fastgpt/global/core/module/runtime/utils';
 import { ChatCompletionRequestMessageRoleEnum } from '@fastgpt/global/core/ai/constants';
@ -25,8 +25,10 @@ import json5 from 'json5';
 import { DispatchFlowResponse } from '../../type';
 import { countGptMessagesTokens } from '@fastgpt/global/common/string/tiktoken';
 import { getNanoid } from '@fastgpt/global/common/string/tools';
+import { AIChatItemType, AIChatItemValueItemType } from '@fastgpt/global/core/chat/type';
+import { GPTMessages2Chats } from '@fastgpt/global/core/chat/adapt';

-type ToolRunResponseType = {
+type FunctionRunResponseType = {
  moduleRunResponse: DispatchFlowResponse;
  functionCallMsg: ChatCompletionFunctionMessageParam;
 }[];
@ -49,6 +51,7 @@ export const runToolWithFunctionCall = async (
    module,
    stream
  } = props;
+  const assistantResponses = response?.assistantResponses || [];

  const functions: ChatCompletionCreateParams.Function[] = toolModules.map((module) => {
    const properties: Record<
@ -161,10 +164,18 @@ export const runToolWithFunctionCall = async (
          startParams
        });

+        const stringToolResponse = (() => {
+          if (typeof moduleRunResponse.toolResponses === 'object') {
+            return JSON.stringify(moduleRunResponse.toolResponses, null, 2);
+          }
+
+          return moduleRunResponse.toolResponses ? String(moduleRunResponse.toolResponses) : 'none';
+        })();
+
        const functionCallMsg: ChatCompletionFunctionMessageParam = {
          role: ChatCompletionRequestMessageRoleEnum.Function,
          name: tool.name,
-          content: JSON.stringify(moduleRunResponse.toolResponses, null, 2)
+          content: stringToolResponse
        };

        if (stream && detail) {
@ -177,7 +188,7 @@ export const runToolWithFunctionCall = async (
                toolName: '',
                toolAvatar: '',
                params: '',
-                response: JSON.stringify(moduleRunResponse.toolResponses, null, 2)
+                response: stringToolResponse
              }
            })
          });
@ -189,7 +200,7 @@ export const runToolWithFunctionCall = async (
        };
      })
    )
-  ).filter(Boolean) as ToolRunResponseType;
+  ).filter(Boolean) as FunctionRunResponseType;

  const flatToolsResponseData = toolsRunResponse.map((item) => item.moduleRunResponse).flat();

@ -204,9 +215,11 @@ export const runToolWithFunctionCall = async (
      ...filterMessages,
      assistantToolMsgParams
    ] as ChatCompletionMessageParam[];
-
    const tokens = countGptMessagesTokens(concatToolMessages, undefined, functions);
-
+    const completeMessages = [
+      ...concatToolMessages,
+      ...toolsRunResponse.map((item) => item?.functionCallMsg)
+    ];
    // console.log(tokens, 'tool');

    if (stream && detail) {
@ -216,33 +229,70 @@ export const runToolWithFunctionCall = async (
      });
    }

+    // tool assistant
+    const toolAssistants = toolsRunResponse
+      .map((item) => {
+        const assistantResponses = item.moduleRunResponse.assistantResponses || [];
+        return assistantResponses;
+      })
+      .flat();
+    // tool node assistant
+    const adaptChatMessages = GPTMessages2Chats(completeMessages);
+    const toolNodeAssistant = adaptChatMessages.pop() as AIChatItemType;
+
+    const toolNodeAssistants = [
+      ...assistantResponses,
+      ...toolAssistants,
+      ...toolNodeAssistant.value
+    ];
+
+    // concat tool responses
+    const dispatchFlowResponse = response
+      ? response.dispatchFlowResponse.concat(flatToolsResponseData)
+      : flatToolsResponseData;
+
+    /* check stop signal */
+    const hasStopSignal = flatToolsResponseData.some(
+      (item) => !!item.flowResponses?.find((item) => item.toolStop)
+    );
+    if (hasStopSignal) {
+      return {
+        dispatchFlowResponse,
+        totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
+        completeMessages: filterMessages,
+        assistantResponses: toolNodeAssistants
+      };
+    }
+
    return runToolWithFunctionCall(
      {
        ...props,
-        messages: [...concatToolMessages, ...toolsRunResponse.map((item) => item?.functionCallMsg)]
+        messages: completeMessages
      },
      {
-        dispatchFlowResponse: response
-          ? response.dispatchFlowResponse.concat(flatToolsResponseData)
-          : flatToolsResponseData,
-        totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens
+        dispatchFlowResponse,
+        totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
+        assistantResponses: toolNodeAssistants
      }
    );
  } else {
    // No tool is invoked, indicating that the process is over
-    const completeMessages = filterMessages.concat({
+    const gptAssistantResponse: ChatCompletionAssistantMessageParam = {
      role: ChatCompletionRequestMessageRoleEnum.Assistant,
      content: answer
-    });
-
+    };
+    const completeMessages = filterMessages.concat(gptAssistantResponse);
    const tokens = countGptMessagesTokens(completeMessages, undefined, functions);
-
    // console.log(tokens, 'response token');

+    // concat tool assistant
+    const toolNodeAssistant = GPTMessages2Chats([gptAssistantResponse])[0] as AIChatItemType;
+
    return {
      dispatchFlowResponse: response?.dispatchFlowResponse || [],
      totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
-      completeMessages
+      completeMessages,
+      assistantResponses: [...assistantResponses, ...toolNodeAssistant.value]
    };
  }
 };
--- a/packages/service/core/workflow/dispatch/agent/runTool/index.ts
+++ b/packages/service/core/workflow/dispatch/agent/runTool/index.ts
@ -4,7 +4,7 @@ import type {
  DispatchNodeResultType,
  RunningModuleItemType
 } from '@fastgpt/global/core/module/runtime/type';
-import { ModelTypeEnum, getLLMModel } from '@fastgpt/service/core/ai/model';
+import { ModelTypeEnum, getLLMModel } from '../../../../ai/model';
 import { getHistories } from '../../utils';
 import { runToolWithToolChoice } from './toolChoice';
 import { DispatchToolModuleProps, ToolModuleItemType } from './type.d';
@ -16,9 +16,12 @@ import {
  getSystemPrompt,
  runtimePrompt2ChatsValue
 } from '@fastgpt/global/core/chat/adapt';
-import { formatModelChars2Points } from '@fastgpt/service/support/wallet/usage/utils';
+import { formatModelChars2Points } from '../../../../../support/wallet/usage/utils';
 import { getHistoryPreview } from '@fastgpt/global/core/chat/utils';
 import { runToolWithFunctionCall } from './functionCall';
+import { runToolWithPromptCall } from './promptCall';
+import { replaceVariable } from '@fastgpt/global/common/string/tools';
+import { Prompt_Tool_Call } from './constants';

 type Response = DispatchNodeResultType<{}>;

@ -72,16 +75,19 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
  ];

  const {
-    dispatchFlowResponse,
+    dispatchFlowResponse, // tool flow response
    totalTokens,
-    completeMessages = []
+    completeMessages = [], // The actual message sent to AI(just save text)
+    assistantResponses = [] // FastGPT system store assistant.value response
  } = await (async () => {
+    const adaptMessages = chats2GPTMessages({ messages, reserveId: false });
+
    if (toolModel.toolChoice) {
      return runToolWithToolChoice({
        ...props,
        toolModules,
        toolModel,
-        messages: chats2GPTMessages({ messages, reserveId: false })
+        messages: adaptMessages
      });
    }
    if (toolModel.functionCall) {
@ -89,14 +95,25 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
        ...props,
        toolModules,
        toolModel,
-        messages: chats2GPTMessages({ messages, reserveId: false })
+        messages: adaptMessages
      });
    }
-    return {
-      dispatchFlowResponse: [],
-      totalTokens: 0,
-      completeMessages: []
-    };
+
+    const lastMessage = adaptMessages[adaptMessages.length - 1];
+    if (typeof lastMessage.content !== 'string') {
+      return Promise.reject('暂时只支持纯文本');
+    }
+
+    lastMessage.content = replaceVariable(Prompt_Tool_Call, {
+      question: userChatInput
+    });
+
+    return runToolWithPromptCall({
+      ...props,
+      toolModules,
+      toolModel,
+      messages: adaptMessages
+    });
  })();

  const { totalPoints, modelName } = formatModelChars2Points({
@ -105,11 +122,6 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
    modelType: ModelTypeEnum.llm
  });

-  const adaptMessages = GPTMessages2Chats(completeMessages);
-  //@ts-ignore
-  const startIndex = adaptMessages.findLastIndex((item) => item.obj === ChatRoleEnum.Human);
-  const assistantResponse = adaptMessages.slice(startIndex + 1);
-
  // flat child tool response
  const childToolResponse = dispatchFlowResponse.map((item) => item.flowResponses).flat();

@ -123,9 +135,7 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
  const flatUsages = dispatchFlowResponse.map((item) => item.flowUsages).flat();

  return {
-    [DispatchNodeResponseKeyEnum.assistantResponses]: assistantResponse
-      .map((item) => item.value)
-      .flat(),
+    [DispatchNodeResponseKeyEnum.assistantResponses]: assistantResponses,
    [DispatchNodeResponseKeyEnum.nodeResponse]: {
      totalPoints: totalPointsUsage,
      toolCallTokens: totalTokens,
--- a/packages/service/core/workflow/dispatch/agent/runTool/promptCall.ts
+++ b/packages/service/core/workflow/dispatch/agent/runTool/promptCall.ts
@ -0,0 +1,385 @@
+import { LLMModelItemType } from '@fastgpt/global/core/ai/model.d';
+import { getAIApi } from '../../../../ai/config';
+import { filterGPTMessageByMaxTokens } from '../../../../chat/utils';
+import {
+  ChatCompletion,
+  StreamChatType,
+  ChatCompletionMessageParam,
+  ChatCompletionAssistantMessageParam
+} from '@fastgpt/global/core/ai/type';
+import { NextApiResponse } from 'next';
+import {
+  responseWrite,
+  responseWriteController,
+  responseWriteNodeStatus
+} from '../../../../../common/response';
+import { SseResponseEventEnum } from '@fastgpt/global/core/module/runtime/constants';
+import { textAdaptGptResponse } from '@fastgpt/global/core/module/runtime/utils';
+import { ChatCompletionRequestMessageRoleEnum } from '@fastgpt/global/core/ai/constants';
+import { dispatchWorkFlow } from '../../index';
+import { DispatchToolModuleProps, RunToolResponse, ToolModuleItemType } from './type.d';
+import json5 from 'json5';
+import { countGptMessagesTokens } from '@fastgpt/global/common/string/tiktoken';
+import { getNanoid, replaceVariable } from '@fastgpt/global/common/string/tools';
+import { AIChatItemType } from '@fastgpt/global/core/chat/type';
+import { GPTMessages2Chats } from '@fastgpt/global/core/chat/adapt';
+
+type FunctionCallCompletion = {
+  id: string;
+  name: string;
+  arguments: string;
+  toolName?: string;
+  toolAvatar?: string;
+};
+
+export const runToolWithPromptCall = async (
+  props: DispatchToolModuleProps & {
+    messages: ChatCompletionMessageParam[];
+    toolModules: ToolModuleItemType[];
+    toolModel: LLMModelItemType;
+  },
+  response?: RunToolResponse
+): Promise<RunToolResponse> => {
+  const {
+    toolModel,
+    toolModules,
+    messages,
+    res,
+    runtimeModules,
+    detail = false,
+    module,
+    stream
+  } = props;
+  const assistantResponses = response?.assistantResponses || [];
+
+  const toolsPrompt = JSON.stringify(
+    toolModules.map((module) => {
+      const properties: Record<
+        string,
+        {
+          type: string;
+          description: string;
+          required?: boolean;
+        }
+      > = {};
+      module.toolParams.forEach((item) => {
+        properties[item.key] = {
+          type: 'string',
+          description: item.toolDescription || ''
+        };
+      });
+
+      return {
+        toolId: module.moduleId,
+        description: module.intro,
+        parameters: {
+          type: 'object',
+          properties,
+          required: module.toolParams.filter((item) => item.required).map((item) => item.key)
+        }
+      };
+    })
+  );
+
+  const lastMessage = messages[messages.length - 1];
+  if (typeof lastMessage.content !== 'string') {
+    return Promise.reject('暂时只支持纯文本');
+  }
+  lastMessage.content = replaceVariable(lastMessage.content, {
+    toolsPrompt
+  });
+
+  const filterMessages = filterGPTMessageByMaxTokens({
+    messages,
+    maxTokens: toolModel.maxContext - 500 // filter token. not response maxToken
+  });
+  // console.log(JSON.stringify(filterMessages, null, 2));
+  /* Run llm */
+  const ai = getAIApi({
+    timeout: 480000
+  });
+  const aiResponse = await ai.chat.completions.create(
+    {
+      ...toolModel?.defaultConfig,
+      model: toolModel.model,
+      temperature: 0,
+      stream,
+      messages: filterMessages
+    },
+    {
+      headers: {
+        Accept: 'application/json, text/plain, */*'
+      }
+    }
+  );
+
+  const answer = await (async () => {
+    if (stream) {
+      const { answer } = await streamResponse({
+        res,
+        detail,
+        toolModules,
+        stream: aiResponse
+      });
+
+      return answer;
+    } else {
+      const result = aiResponse as ChatCompletion;
+
+      return result.choices?.[0]?.message?.content || '';
+    }
+  })();
+
+  const parseAnswerResult = parseAnswer(answer);
+  // console.log(answer, '==11==');
+  // No tools
+  if (typeof parseAnswerResult === 'string') {
+    // No tool is invoked, indicating that the process is over
+    const gptAssistantResponse: ChatCompletionAssistantMessageParam = {
+      role: ChatCompletionRequestMessageRoleEnum.Assistant,
+      content: parseAnswerResult
+    };
+    const completeMessages = filterMessages.concat(gptAssistantResponse);
+    const tokens = countGptMessagesTokens(completeMessages, undefined);
+    // console.log(tokens, 'response token');
+
+    // concat tool assistant
+    const toolNodeAssistant = GPTMessages2Chats([gptAssistantResponse])[0] as AIChatItemType;
+
+    return {
+      dispatchFlowResponse: response?.dispatchFlowResponse || [],
+      totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
+      completeMessages,
+      assistantResponses: [...assistantResponses, ...toolNodeAssistant.value]
+    };
+  }
+
+  // Run the selected tool.
+  const toolsRunResponse = await (async () => {
+    if (!parseAnswerResult) return Promise.reject('tool run error');
+
+    const toolModule = toolModules.find((module) => module.moduleId === parseAnswerResult.name);
+    if (!toolModule) return Promise.reject('tool not found');
+
+    parseAnswerResult.toolName = toolModule.name;
+    parseAnswerResult.toolAvatar = toolModule.avatar;
+
+    // SSE response to client
+    if (stream && detail) {
+      responseWrite({
+        res,
+        event: SseResponseEventEnum.toolCall,
+        data: JSON.stringify({
+          tool: {
+            id: parseAnswerResult.id,
+            toolName: toolModule.name,
+            toolAvatar: toolModule.avatar,
+            functionName: parseAnswerResult.name,
+            params: parseAnswerResult.arguments,
+            response: ''
+          }
+        })
+      });
+    }
+
+    // run tool flow
+    const startParams = (() => {
+      try {
+        return json5.parse(parseAnswerResult.arguments);
+      } catch (error) {
+        return {};
+      }
+    })();
+
+    const moduleRunResponse = await dispatchWorkFlow({
+      ...props,
+      runtimeModules: runtimeModules.map((module) => ({
+        ...module,
+        isEntry: module.moduleId === toolModule.moduleId
+      })),
+      startParams
+    });
+
+    const stringToolResponse = (() => {
+      if (typeof moduleRunResponse.toolResponses === 'object') {
+        return JSON.stringify(moduleRunResponse.toolResponses, null, 2);
+      }
+
+      return moduleRunResponse.toolResponses ? String(moduleRunResponse.toolResponses) : 'none';
+    })();
+
+    if (stream && detail) {
+      responseWrite({
+        res,
+        event: SseResponseEventEnum.toolResponse,
+        data: JSON.stringify({
+          tool: {
+            id: parseAnswerResult.id,
+            toolName: '',
+            toolAvatar: '',
+            params: '',
+            response: stringToolResponse
+          }
+        })
+      });
+    }
+
+    return {
+      moduleRunResponse,
+      toolResponsePrompt: stringToolResponse
+    };
+  })();
+
+  if (stream && detail) {
+    responseWriteNodeStatus({
+      res,
+      name: module.name
+    });
+  }
+
+  // 合并工具调用的结果，使用 functionCall 格式存储。
+  const assistantToolMsgParams: ChatCompletionAssistantMessageParam = {
+    role: ChatCompletionRequestMessageRoleEnum.Assistant,
+    function_call: parseAnswerResult
+  };
+  const concatToolMessages = [
+    ...filterMessages,
+    assistantToolMsgParams
+  ] as ChatCompletionMessageParam[];
+  const tokens = countGptMessagesTokens(concatToolMessages, undefined);
+  const completeMessages: ChatCompletionMessageParam[] = [
+    ...concatToolMessages,
+    {
+      role: ChatCompletionRequestMessageRoleEnum.Function,
+      name: parseAnswerResult.name,
+      content: toolsRunResponse.toolResponsePrompt
+    }
+  ];
+
+  // tool assistant
+  const toolAssistants = toolsRunResponse.moduleRunResponse.assistantResponses || [];
+  // tool node assistant
+  const adaptChatMessages = GPTMessages2Chats(completeMessages);
+  const toolNodeAssistant = adaptChatMessages.pop() as AIChatItemType;
+
+  const toolNodeAssistants = [...assistantResponses, ...toolAssistants, ...toolNodeAssistant.value];
+
+  const dispatchFlowResponse = response
+    ? response.dispatchFlowResponse.concat(toolsRunResponse.moduleRunResponse)
+    : [toolsRunResponse.moduleRunResponse];
+
+  // get the next user prompt
+  lastMessage.content += `${answer}
+TOOL_RESPONSE: ${toolsRunResponse.toolResponsePrompt}
+ANSWER: `;
+
+  /* check stop signal */
+  const hasStopSignal = toolsRunResponse.moduleRunResponse.flowResponses.some(
+    (item) => !!item.toolStop
+  );
+  if (hasStopSignal) {
+    return {
+      dispatchFlowResponse,
+      totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
+      completeMessages: filterMessages,
+      assistantResponses: toolNodeAssistants
+    };
+  }
+
+  return runToolWithPromptCall(
+    {
+      ...props,
+      messages
+    },
+    {
+      dispatchFlowResponse,
+      totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
+      assistantResponses: toolNodeAssistants
+    }
+  );
+};
+
+async function streamResponse({
+  res,
+  detail,
+  stream
+}: {
+  res: NextApiResponse;
+  detail: boolean;
+  toolModules: ToolModuleItemType[];
+  stream: StreamChatType;
+}) {
+  const write = responseWriteController({
+    res,
+    readStream: stream
+  });
+
+  let startResponseWrite = false;
+  let textAnswer = '';
+
+  for await (const part of stream) {
+    if (res.closed) {
+      stream.controller?.abort();
+      break;
+    }
+
+    const responseChoice = part.choices?.[0]?.delta;
+    if (responseChoice.content) {
+      const content = responseChoice?.content || '';
+      textAnswer += content;
+
+      if (startResponseWrite) {
+        responseWrite({
+          write,
+          event: detail ? SseResponseEventEnum.answer : undefined,
+          data: textAdaptGptResponse({
+            text: content
+          })
+        });
+      } else if (textAnswer.length >= 3) {
+        textAnswer = textAnswer.trim();
+        if (textAnswer.startsWith('0')) {
+          startResponseWrite = true;
+          // find first : index
+          const firstIndex = textAnswer.indexOf(':');
+          textAnswer = textAnswer.substring(firstIndex + 1).trim();
+          responseWrite({
+            write,
+            event: detail ? SseResponseEventEnum.answer : undefined,
+            data: textAdaptGptResponse({
+              text: textAnswer
+            })
+          });
+        }
+      }
+    }
+  }
+
+  if (!textAnswer) {
+    return Promise.reject('LLM api response empty');
+  }
+  // console.log(textAnswer, '---===');
+  return { answer: textAnswer.trim() };
+}
+
+const parseAnswer = (str: string): FunctionCallCompletion | string => {
+  // 首先，使用正则表达式提取TOOL_ID和TOOL_ARGUMENTS
+  const prefix = '1:';
+  str = str.trim();
+  if (str.startsWith(prefix)) {
+    const toolString = str.substring(prefix.length).trim();
+
+    try {
+      const toolCall = json5.parse(toolString);
+      return {
+        id: getNanoid(),
+        name: toolCall.toolId,
+        arguments: JSON.stringify(toolCall.arguments)
+      };
+    } catch (error) {
+      return str;
+    }
+  } else {
+    return str;
+  }
+};
--- a/Show More
+++ b/Show More