Agno Docs MCP Server
Provides access to Agno framework documentation for AI agents, enabling search and retrieval of SDK references, API endpoints, code examples, and integration guides through MCP-compatible tools.
README
Agno Docs MCP Server
A Model Context Protocol (MCP) server that provides access to Agno framework documentation. Enables developers to easily access Agno docs through coding agents like Claude Code, Cursor, and other MCP-compatible tools.
Quick Start (New Laptop Setup)
Prerequisites
- Python 3.10 or higher
- Git
- Access to the
agno-docsrepository (for documentation source)
Step 1: Clone the Repository
git clone https://github.com/agno-agi/agno-docs-mcp
cd agno-docs-mcp
Step 2: Create Virtual Environment
python3 -m venv .venv
source .venv/bin/activate # On Windows: .venv\Scripts\activate
Step 3: Install Dependencies
# Upgrade pip first (required for pyproject.toml editable installs)
pip install --upgrade pip
# Install the package in editable mode with dev dependencies
pip install -e ".[dev]"
Step 4: Clone Agno Docs (if not already available)
# Clone the agno-docs repository (adjust path as needed)
git clone https://github.com/agno-agi/agno-docs ~/Work/agno-docs
Step 5: Prepare Documentation
# Set the path to your agno-docs repository
export AGNO_DOCS_PATH=~/Work/agno-docs
# Run the preparation script to copy and index docs
python -m agno_docs_mcp.prepare
You should see output like:
Preparing Agno documentation...
Source: /Users/you/Work/agno-docs
Output: /Users/you/Work/agno-docs-mcp/.docs
Copying documentation files...
Copied 596 files from basics/
Copied 148 files from reference/
Copied 74 files from reference-api/
...
Copied OpenAPI spec (openapi.json)
Building search index...
Indexed 1851 files in 10 categories
Done!
Step 6: Configure Your MCP Client
Claude Code
Add to ~/.claude.json or your project's .claude/settings.json:
{
"mcpServers": {
"agno-docs": {
"command": "/path/to/agno-docs-mcp/.venv/bin/python",
"args": ["-m", "agno_docs_mcp"]
}
}
}
Cursor
Add to your Cursor MCP settings (.cursor/mcp.json):
{
"mcpServers": {
"agno-docs": {
"command": "/path/to/agno-docs-mcp/.venv/bin/python",
"args": ["-m", "agno_docs_mcp"]
}
}
}
Using HTTP Transport (Remote Access)
{
"mcpServers": {
"agno-docs": {
"url": "http://localhost:8000/mcp",
"transport": "streamable-http"
}
}
}
Step 7: Verify Installation
# Test the server directly
python -c "from agno_docs_mcp.tools.api import agno_api; print(agno_api('memory')[:200])"
# Or start the HTTP server
python -m agno_docs_mcp --transport http --port 8000
# In another terminal, test with curl
curl -X POST http://localhost:8000/mcp \
-H "Content-Type: application/json" \
-H "Accept: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'
Features
-
7 Specialized Tools for navigating Agno documentation:
agno_docs- SDK conceptual documentation (agents, teams, workflows, basics)agno_reference- SDK class and method referenceagno_examples- Code snippets and usage examplesagno_integrations- Database, VectorDB, and model provider guidesagno_agentos- AgentOS runtime documentationagno_api- REST API endpoints from OpenAPI specagno_migration- Migration guides, FAQs, and troubleshooting
-
OpenAPI Integration - Parses the OpenAPI spec for accurate REST endpoint docs
-
Keyword-based search across all documentation
-
Path-based navigation for exploring docs structure
-
Offline support - docs are preprocessed and bundled locally
-
Multiple transports - stdio (local) and HTTP (remote) support
Running the Server
Local CLI Mode (stdio) - Default
# Activate venv first
source .venv/bin/activate
# Run with stdio transport (for Claude Code, Cursor)
python -m agno_docs_mcp
HTTP Server Mode
# Run HTTP server on port 8000
python -m agno_docs_mcp --transport http --port 8000
# Server available at http://localhost:8000/mcp
# Health check at http://localhost:8000/health
Production Deployment
# Using uvicorn directly
uvicorn agno_docs_mcp.app:app --host 0.0.0.0 --port 8000
# Using Docker
docker build -t agno-docs-mcp .
docker run -p 8000:8000 agno-docs-mcp
Tools Reference
agno_api (NEW - REST API Endpoints)
Get AgentOS REST API endpoints from the OpenAPI specification.
agno_api(resource="memory")
Parameters:
resource(str) - API resource: memory, agents, teams, workflows, sessions, knowledge, evals, traces, metrics, database, playground
Use for: REST API endpoints, HTTP methods, request/response schemas
agno_docs
Get SDK conceptual documentation and guides for writing agent code.
agno_docs(path="basics/agents/")
Parameters:
path(str) - Documentation path (e.g., "basics/", "basics/agents/", "basics/memory/")
Use for: How to write Python code with Agno SDK
agno_reference
Get SDK class and method reference (parameters, signatures, options).
agno_reference(topic="agents")
Parameters:
topic(str) - Reference topic: agents, teams, workflows, tools, models, memory, knowledge, storage, hooks, compression, reasoning, agent-os
Use for: Agent() constructor parameters, method signatures, configuration options
agno_examples
Get SDK code examples for building agents.
agno_examples(category="agents")
Parameters:
category(str) - Category: agents, teams, workflows, tools, memory, knowledge, models, database, evals, guardrails, hitl, multimodal, reasoning, sessions, tracing
agno_integrations
Get integration documentation for databases, vector stores, and models.
agno_integrations(integration_type="database", name="postgres")
Parameters:
integration_type(str) - Type: database, vectordb, models, toolkitsname(str, optional) - Specific integration name
agno_agentos
Get AgentOS runtime and deployment documentation.
agno_agentos(path="features/memories")
Parameters:
path(str) - Path within AgentOS docs (e.g., "api/", "features/", "security/")
Use for: Deployment docs, runtime features, authentication, middleware
agno_migration
Get migration guides and FAQ documentation.
agno_migration(topic="v2-migration")
Parameters:
topic(str) - Migration topic or FAQ topic
Tool Selection Guide
| Question Type | Use This Tool |
|---|---|
| "What REST API endpoints for memory?" | agno_api("memory") |
| "How to create an agent in Python?" | agno_docs("basics/agents/") |
| "What parameters does Agent() accept?" | agno_reference("agents") |
| "Show me agent code examples" | agno_examples("agents") |
| "How to connect to Postgres?" | agno_integrations("database", "postgres") |
| "How to deploy agents?" | agno_agentos("api/") |
| "How to migrate from v1?" | agno_migration("v2-migration") |
Project Structure
agno-docs-mcp/
├── pyproject.toml
├── README.md
├── .venv/ # Virtual environment
├── src/
│ └── agno_docs_mcp/
│ ├── __init__.py
│ ├── __main__.py
│ ├── server.py # Main MCP server with tool registration
│ ├── app.py # FastAPI/ASGI app for HTTP
│ ├── tools/ # Tool implementations
│ │ ├── docs.py
│ │ ├── reference.py
│ │ ├── examples.py
│ │ ├── integrations.py
│ │ ├── agentos.py
│ │ ├── api.py # NEW: OpenAPI parser
│ │ └── migration.py
│ ├── utils/ # Utility modules
│ │ ├── search.py
│ │ ├── paths.py
│ │ └── content.py
│ └── prepare/ # Doc preparation
│ └── prepare_docs.py
├── .docs/ # Preprocessed docs (generated)
│ ├── raw/ # Copied MDX files
│ ├── snippets/ # Snippet files
│ ├── index.json # Search index
│ └── openapi.json # OpenAPI spec
└── tests/
Updating Documentation
When the agno-docs repository is updated:
# Pull latest docs
cd ~/Work/agno-docs
git pull
# Re-run preparation
cd ~/Work/agno-docs-mcp
source .venv/bin/activate
python -m agno_docs_mcp.prepare
Troubleshooting
"OpenAPI specification not found"
Run the prepare script:
export AGNO_DOCS_PATH=~/Work/agno-docs
python -m agno_docs_mcp.prepare
"Module not found" errors
Make sure you're using the virtual environment:
source .venv/bin/activate
which python # Should show .venv/bin/python
MCP client not connecting
-
Check the path in your MCP config points to the venv Python:
"command": "/absolute/path/to/agno-docs-mcp/.venv/bin/python" -
Test the server manually:
/path/to/agno-docs-mcp/.venv/bin/python -m agno_docs_mcp
License
MIT
Links
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。