ima-mcp-server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| IMA_API_KEY | No | Your IMA API key (from https://ima.qq.com/agent-interface) | |
| IMA_CLIENT_ID | No | Your IMA client ID (from https://ima.qq.com/agent-interface) | |
| IMA_OPENAPI_APIKEY | No | Alternative environment variable name for IMA API key (ima-skill compatibility) | |
| IMA_OPENAPI_CLIENTID | No | Alternative environment variable name for IMA client ID (ima-skill compatibility) |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_media_infoA | 获取知识库条目的原文内容或下载链接。传入 media_id(来自 ask_knowledge_base 返回的 references 字段)。对微信文章/网页/文件类,返回可访问 URL;对笔记类(media_id 以 note_ 开头),会自动拉取并返回笔记正文(note_content 字段,纯文本)。 |
| get_note_contentA | 读取笔记正文(纯文本)。传入 note_id(笔记的数字 ID,即 get_media_info 返回的 notebook_id,或笔记搜索接口返回的 doc_id)。这是读取 IMA 笔记原文的专用接口。 |
| get_addable_knowledge_base_listA | 获取当前用户有权限添加内容的知识库列表。仅在未指定目标知识库时使用。 |
| import_urlsB | 将网页或微信文章 URL 添加到知识库(1-10 个 URL)。添加到根目录时省略 folder_id。 |
| create_noteA | 创建一篇 IMA 笔记。content 支持 Markdown 格式(content_format=1)。创建后可通过 add_note_to_knowledge_base 将笔记添加到知识库。folder_name 用于指定笔记在 IMA 笔记中的存放文件夹(不是知识库文件夹)。 |
| add_note_to_knowledge_baseA | 将已有的 IMA 笔记添加到知识库。需要先通过 create_note 创建笔记获取 note_id,或将 get_media_info 返回的 notebook_id 作为 note_id 传入。 |
| check_repeated_namesA | 上传文件前检查知识库中是否已存在同名文件。返回每个文件的重复状态。仅适用于文件类型(PDF/Word/PPT/Excel/图片等),不适用于网页和笔记。 |
| upload_fileA | 上传本地文件到知识库。支持 PDF、Word(.doc/.docx)、PPT(.ppt/.pptx)、Excel(.xls/.xlsx/.csv)、Markdown(.md)、图片(PNG/JPG/WebP)、TXT、XMind、音频(MP3/M4A/WAV/AAC)。会自动检查文件名重复、创建媒体、上传到 COS、注册到知识库。 |
| ask_knowledge_baseA | 向知识库提问,使用 IMA 完整的 RAG 搜索引擎(语义检索 + 知识图谱 + LLM 生成回答)。这是与 IMA 网页端完全一致的问答能力,能深入检索文档全文内容。需要配置 IMA_COOKIE 环境变量(从浏览器开发者工具获取)。⚠️ knowledge_base_id 请使用 list_knowledge_bases_full 返回的 cookie_knowledge_base_id 字段。典型用法:'上周会议的主要结论是什么?'、'总结知识库里关于 X 的所有内容'、'这些文档中的关键数据是什么?' |
| list_knowledge_bases_fullA | 获取完整的知识库列表(含订阅的共享知识库),同时返回两套 ID:digit_id(纯数字)供 ask_knowledge_base 使用,openapi_knowledge_base_id(base64url 格式)供所有其他知识库工具使用。这是获取知识库 ID 的唯一入口,需要配置 IMA_COOKIE。 |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 10 tools
Most tools have clear distinct purposes, but get_media_info and get_note_content both retrieve note content, which could cause confusion. However, they accept different ID types (media_id vs note_id) and are described with explicit usage contexts, reducing ambiguity.
All tool names follow a consistent snake_case verb_noun pattern (get, import, create, add, check, upload, list, ask). There is no mixing of naming conventions or inconsistent verb styles.
With 10 tools, the server is well-scoped for an IMA knowledge base assistant. It covers listing knowledge bases, adding content (URLs, files, notes), retrieving content, and querying without being overwhelming or sparse.
The tool set covers core workflows: list KBs, add content via URLs/files/notes, get media/note content, and ask questions. However, it lacks update/delete operations for content or knowledge bases, and there is no direct search or browse tool to list KB entries without using ask_knowledge_base. References are only obtainable through ask_knowledge_base, creating a potential dead end.