CouchDB MCP Server
Provides tools for interacting with CouchDB databases, enabling document management, Mango search queries, and database administration. It allows Claude Code and other MCP clients to perform CRUD operations and manage indexes on local or remote CouchDB instances.
README
CouchDB MCP Server
A Model Context Protocol (MCP) server that provides tools for interacting with CouchDB databases. This server enables Claude Code and other MCP clients to perform database operations, document management, and search queries on CouchDB instances.
Features
Database Operations
- List Databases: View all databases on the CouchDB server
- Create Database: Create new databases
- Delete Database: Remove databases
Document Operations
- Create Document: Insert new documents with optional custom IDs
- Get Document: Retrieve documents by ID
- Update Document: Modify existing documents
- Delete Document: Remove documents
- List Documents: View all documents in a database with optional full content
Search & Indexing
- Search Documents: Query documents using CouchDB Mango queries with pagination support
- Create Index: Create indexes for better query performance
- List Indexes: View all indexes in a database
Requirements
- Python 3.10 or higher
- CouchDB server (local or remote)
- Claude Code CLI
Installation
-
Clone or download this repository to your local machine.
-
Install dependencies using uv (recommended):
uv sync
Or using pip:
pip install -e .
CouchDB Setup
Make sure you have CouchDB installed and running. You can:
-
Install CouchDB locally:
- macOS:
brew install couchdb - Ubuntu/Debian:
sudo apt-get install couchdb - Or download from couchdb.apache.org
- macOS:
-
Use Docker:
docker run -d -p 5984:5984 --name couchdb \ -e COUCHDB_USER=admin \ -e COUCHDB_PASSWORD=password \ couchdb:latest -
Access CouchDB:
- Default URL:
http://localhost:5984 - Web UI (Fauxton):
http://localhost:5984/_utils
- Default URL:
Adding to Claude Code
To use this MCP server with Claude Code, you need to add it to your Claude Code configuration file.
Configuration File Location
The Claude Code configuration file is located at:
- macOS/Linux:
~/.config/claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
Configuration Steps
- Open or create the configuration file:
# macOS/Linux
mkdir -p ~/.config/claude
nano ~/.config/claude/claude_desktop_config.json
- Add the CouchDB MCP server to the
mcpServerssection:
Using uv (recommended):
{
"mcpServers": {
"couchdb": {
"command": "uv",
"args": [
"run",
"--directory", "/path/to/couchdb_mcp",
"python", "couchdb_mcp_server.py"
],
"env": {
"COUCHDB_URL": "http://localhost:5984"
}
}
}
}
Using python directly (requires manual dependency install):
{
"mcpServers": {
"couchdb": {
"command": "python",
"args": [
"/path/to/couchdb_mcp_server.py"
],
"env": {
"COUCHDB_URL": "http://localhost:5984"
}
}
}
}
- For authenticated CouchDB instances, include credentials in the URL:
{
"mcpServers": {
"couchdb": {
"command": "uv",
"args": [
"run",
"--directory", "/path/to/couchdb_mcp",
"python", "couchdb_mcp_server.py"
],
"env": {
"COUCHDB_URL": "http://admin:password@localhost:5984"
}
}
}
}
- For remote CouchDB servers:
{
"mcpServers": {
"couchdb": {
"command": "uv",
"args": [
"run",
"--directory", "/path/to/couchdb_mcp",
"python", "couchdb_mcp_server.py"
],
"env": {
"COUCHDB_URL": "https://username:password@your-server.com:5984"
}
}
}
}
- Save the configuration file and restart Claude Code.
Usage Examples
Once configured, you can use Claude Code to interact with your CouchDB instance. Here are some example requests:
Database Operations
Create a new database called "users"
List all databases
Delete the database named "test_db"
Document Operations
Create a document in the users database with data: {"name": "John Doe", "email": "john@example.com"}
Get the document with ID "user123" from the users database
Update document user123 in users database with new email address
List all documents in the users database
Search Operations
Search the users database for all documents where name equals "John Doe"
Search for documents in the products database where price is greater than 100
Tool Reference
couchdb_list_databases
Lists all databases on the CouchDB server.
Parameters: None
couchdb_create_database
Creates a new database.
Parameters:
name(string, required): Name of the database to create
couchdb_delete_database
Deletes a database.
Parameters:
name(string, required): Name of the database to delete
couchdb_create_document
Creates a new document in a database.
Parameters:
database(string, required): Name of the databasedocument(object, required): Document data as JSON objectdoc_id(string, optional): Document ID (auto-generated if not provided)
couchdb_get_document
Retrieves a document from a database.
Parameters:
database(string, required): Name of the databasedoc_id(string, required): Document ID
couchdb_update_document
Updates an existing document.
Parameters:
database(string, required): Name of the databasedoc_id(string, required): Document IDdocument(object, required): Updated document data (must include_rev)
couchdb_delete_document
Deletes a document from a database.
Parameters:
database(string, required): Name of the databasedoc_id(string, required): Document IDrev(string, required): Document revision (_rev)
couchdb_search_documents
Searches for documents using Mango queries.
Parameters:
database(string, required): Name of the databasequery(object, required): Mango query selector (e.g.,{"name": "John"})limit(integer, optional): Maximum number of documents to return (default: 25)skip(integer, optional): Number of documents to skip (default: 0)
couchdb_list_documents
Lists all documents in a database.
Parameters:
database(string, required): Name of the databaselimit(integer, optional): Maximum number of documents to returninclude_docs(boolean, optional): Include full document content (default: false)
couchdb_create_index
Creates an index to dramatically improve Mango query performance.
Parameters:
database(string, required): Name of the databasefields(array, required): Fields to index (e.g.,["type", "name"])index_name(string, optional): Name for the index
Note: While indexes are optional, they are highly recommended. Without indexes, CouchDB scans all documents which can be very slow on large databases.
couchdb_list_indexes
Lists all indexes in a database.
Parameters:
database(string, required): Name of the database
About Indexes and Mango Queries
CouchDB's Mango query system (couchdb_search_documents) does not require indexes, but they are strongly recommended for performance.
According to the CouchDB documentation:
- Without an index, CouchDB falls back to scanning all documents (
_all_docs) - This works but "can be arbitrarily slow" on large databases
- Creating indexes dramatically improves query performance
When to create indexes:
- Your database has many documents (>1000)
- Queries are slow
- You frequently query the same fields
When you can skip indexes:
- Small databases (<100 documents)
- Infrequent queries
- You don't mind slower response times
Example:
1. Try searching: "Search omnibot database where type equals 'name'"
2. If it works but is slow, create an index: "Create an index on the 'type' field in omnibot database"
3. Future queries will be much faster
Troubleshooting
Connection Issues
If you see connection errors:
- Verify CouchDB is running:
curl http://localhost:5984 - Check the URL in your configuration
- Verify credentials if using authentication
- Check firewall settings for remote connections
Permission Errors
If you see permission errors:
- Ensure the user has appropriate CouchDB permissions
- Check that the database exists before performing document operations
- Verify document revisions (
_rev) when updating or deleting
Search Returns No Results
If you're searching for documents but getting no results:
- Verify documents exist: Use
couchdb_list_documentswithinclude_docs: trueto see actual document structure - Check field names match exactly: Field names are case-sensitive (
"Type"≠"type") - Verify field values match: Values are also case-sensitive (
"Name"≠"name") - Check the field exists: Queries only match documents where the field is present
- Consider creating an index: While not required, indexes ensure queries work reliably and quickly
Tool Not Found
If Claude Code doesn't recognize the CouchDB tools:
- Verify the configuration file path is correct
- Ensure the
--directorypath (uv) or Python script path is absolute, not relative - Restart Claude Code after configuration changes
- Check that dependencies are installed (
uv syncorpip install -e .)
Mango Query Examples
Mango queries use MongoDB-style selectors:
Equality:
{"name": "John Doe"}
Comparison:
{"age": {"$gt": 18}}
Multiple conditions:
{"$and": [{"age": {"$gte": 18}}, {"status": "active"}]}
Pattern matching:
{"email": {"$regex": ".*@example\\.com$"}}
License
MIT License
Contributing
Contributions are welcome! Please feel free to submit issues or pull requests.
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。