Keboola API Documentation MCP Server
Give your AI agents instant access to Keboola API documentation. Search endpoints, get parameter details, and generate curl examples—all through MCP tools.
README
Keboola API Documentation MCP Server
Give your AI agents instant access to Keboola API documentation. Search endpoints, get parameter details, and generate curl examples—all through MCP tools.
Overview
This MCP (Model Context Protocol) server provides AI agents with searchable access to all Keboola API documentation. Instead of manually searching through API docs, your agents can query endpoints, explore sections, and get ready-to-use examples.
Indexed APIs:
- Storage API (199 endpoints)
- Management API (129 endpoints)
- Stream Service (41 endpoints)
- Templates Service (23 endpoints)
- Orchestrator API (15 endpoints)
- Metastore Service (14 endpoints)
- Query Service (10 endpoints)
- Generic Extractor API (8 endpoints)
- AppsProxy Service (7 endpoints)
Features
- Search endpoints by keyword, API name, or HTTP method
- Browse by section (Tables, Buckets, Jobs, etc.)
- Get full details including parameters, request/response examples
- Generate curl examples for any endpoint
- Auto-update documentation from GitHub sources
Quick Start
Installation
# Clone the repository
git clone https://github.com/keboola/keboola-api-documentation-mcp-server
cd keboola-api-documentation-mcp-server
# Install dependencies
poetry install
# Update documentation from GitHub
poetry run keboola-docs update
MCP Client Setup
Claude Desktop Configuration
- Open Claude Desktop
- Go to Claude (menu) → Settings → Developer → Edit Config
- Add the following configuration:
{
"mcpServers": {
"keboola-docs": {
"command": "poetry",
"args": ["--directory", "/path/to/keboola-api-documentation-mcp-server", "run", "python", "-m", "keboola_docs_mcp.server"]
}
}
}
Config file locations:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
Claude Code Configuration
Add the MCP server using the Claude Code CLI:
claude mcp add keboola-docs -- poetry --directory /path/to/keboola-api-documentation-mcp-server run python -m keboola_docs_mcp.server
Or add manually to your Claude Code settings:
{
"mcpServers": {
"keboola-docs": {
"command": "poetry",
"args": ["--directory", "/path/to/keboola-api-documentation-mcp-server", "run", "python", "-m", "keboola_docs_mcp.server"]
}
}
}
Cursor Configuration
- Go to Settings → MCP
- Click "+ Add new global MCP Server"
- Configure with these settings:
{
"mcpServers": {
"keboola-docs": {
"command": "poetry",
"args": ["--directory", "/path/to/keboola-api-documentation-mcp-server", "run", "python", "-m", "keboola_docs_mcp.server"]
}
}
}
Windsurf Configuration
Add to your Windsurf MCP settings:
{
"mcpServers": {
"keboola-docs": {
"command": "poetry",
"args": ["--directory", "/path/to/keboola-api-documentation-mcp-server", "run", "python", "-m", "keboola_docs_mcp.server"]
}
}
}
Using uvx (Alternative)
If you prefer using uvx instead of Poetry:
{
"mcpServers": {
"keboola-docs": {
"command": "uvx",
"args": ["--from", "git+https://github.com/keboola/keboola-api-documentation-mcp-server", "python", "-m", "keboola_docs_mcp.server"]
}
}
}
Available MCP Tools
| Tool | Description |
|---|---|
list_apis() |
List all available Keboola APIs with endpoint counts |
search_endpoints(query, api_filter?, method_filter?, limit?) |
Search for endpoints by keyword |
get_endpoint_details(api_name, path, method) |
Get full endpoint documentation |
get_api_section(api_name, section_name?) |
Get all endpoints in a section |
list_sections(api_name) |
List all sections in an API |
get_request_example(api_name, path, method) |
Generate curl example |
Example Queries
Once configured, you can ask your AI agent:
Searching:
- "Find endpoints for creating tables in Storage API"
- "Search for token-related endpoints"
- "What POST endpoints exist in the Management API?"
Exploring:
- "List all available Keboola APIs"
- "What sections does the Storage API have?"
- "Show me all endpoints in the Tables section"
Details:
- "Get details for POST /v2/storage/buckets"
- "Show me a curl example for creating a table"
- "What parameters does the token verification endpoint need?"
CLI Commands
Update Documentation
Fetch the latest API documentation from GitHub:
# Update all docs
poetry run keboola-docs update
# Update specific API
poetry run keboola-docs update --api storage
# Dry run (check for updates without downloading)
poetry run keboola-docs update --dry-run
List Sources
poetry run keboola-docs list-sources
Run Server Manually
poetry run python -m keboola_docs_mcp.server
CI/CD Integration
GitHub Actions Workflow
The repository includes a workflow that automatically updates documentation daily. It runs at 6:00 UTC and commits any changes.
To trigger manually:
- Go to Actions → Update API Documentation
- Click Run workflow
Custom CI Integration
Add to your workflow:
- name: Set up Python
uses: actions/setup-python@v5
with:
python-version: '3.12'
- name: Install Poetry
uses: snok/install-poetry@v1
- name: Install dependencies
run: poetry install
- name: Update API documentation
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
run: poetry run keboola-docs update
- name: Commit changes
run: |
git add docs/
git diff --staged --quiet || git commit -m "chore: update API docs"
git push
Configuration
sources.yaml
Documentation sources are configured in sources.yaml. Each source specifies:
name: API display nameurl: GitHub raw URL to fetch fromoutput: Local file pathformat:apib(API Blueprint) oropenapidescription: Brief descriptionauth_header: Authentication header name (optional)base_url: API base URL (optional)
Example:
sources:
- name: Storage API
url: https://raw.githubusercontent.com/keboola/storage-api-php-client/master/apiary.apib
output: docs/apiary/storage-api.apib
format: apib
description: Core data storage API
auth_header: X-StorageApi-Token
base_url: https://connection.keboola.com
Private Repositories
The updater automatically uses GITHUB_TOKEN or GH_TOKEN environment variables to access private repositories. In CI, this is provided by GitHub Actions automatically.
Development
Running Tests
poetry run pytest
Project Structure
keboola-api-documentation-mcp-server/
├── docs/
│ ├── apiary/ # API Blueprint documentation
│ └── openapi/ # OpenAPI/Swagger specs
├── src/keboola_docs_mcp/
│ ├── server.py # MCP server with tools
│ ├── cli.py # CLI commands
│ ├── updater.py # Documentation fetcher
│ ├── index.py # Search index
│ ├── models.py # Data models
│ └── parsers/ # API Blueprint & OpenAPI parsers
├── tests/
├── sources.yaml # Documentation sources config
└── pyproject.toml
Troubleshooting
| Issue | Solution |
|---|---|
| No endpoints found | Run poetry run keboola-docs update to fetch docs |
| HTTP 404 errors | Check GITHUB_TOKEN is set for private repos |
| Server won't start | Ensure Poetry dependencies are installed |
| Search returns nothing | Try broader search terms or check API name |
Resources
License
MIT
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。