# PipeCD Docs MCP Server

A local MCP server for searching official PipeCD docs.

## Overview

This server clones the official [PipeCD](https://github.com/pipe-cd/pipecd) docs from GitHub and provides simple full-text search and document retrieval APIs via the MCP protocol.  
Documentation is cloned into a temporary directory, and Markdown files are indexed by extracting their titles and content.

## Usage

### Prerequisites

1. Create a PAT with `read:packages` scope. [Details](https://docs.github.com/en/packages/working-with-a-github-packages-registry/working-with-the-npm-registry)

2. Add the following config to your `.npmrc` file.

```npmrc
@pipe-cd:registry=https://npm.pkg.github.com
//npm.pkg.github.com/:_authToken=YOUR_TOKEN
```

### Cursor

### Quick Install with Cursor Deeplink

For Cursor users, you can install the MCP server with a single click using the deeplink below:

[![Install MCP Server](https://cursor.com/deeplink/mcp-install-dark.svg)](https://cursor.com/en-US/install-mcp?name=pipe-cd.docs-mcp-server&config=eyJ0eXBlIjoic3RkaW8iLCJjb21tYW5kIjoibnB4IEBwaXBlLWNkL2RvY3MtbWNwLXNlcnZlckBsYXRlc3QifQ%3D%3D)

This will automatically configure the MCP server in your Cursor settings. After clicking the link, Cursor will prompt you to install the server.

### Install Using npx

Add the following config to your mcp.json.

```json
{
  "mcpServers": {
    "pipe-cd.docs-mcp-server": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "@pipe-cd/docs-mcp-server@latest"
      ], 
    }
  }
}
```

<!-- TODO: Add Claude Desktop etc. -->

## Tools

### search_docs

Executes a full-text search on PipeCD docs.

- Parameters:
  - `query`: Search keywords (space-separated, AND search)
  - `offset`: Start position for results
  - `limit`: Number of results to return (default: 20)

### read_docs

Returns the content of the specified page.

- Parameters:
  - `path`: Relative path of the document (after "docs/content/en/")

## Implementation Notes

- Uses sparse-checkout to minimize clone size and speed up the process.
- Titles are extracted from the Markdown front matter.
- The search logic of `search_docs` is so simple for now.


## Contributing

### Code of Conduct

PipeCD follows the [CNCF Code of Conduct](https://github.com/cncf/foundation/blob/master/code-of-conduct.md). Please read it to understand which actions are acceptable and which are not.

### Get Involved

- Slack: `#pipecd` channel on [CNCF Slack](https://cloud-native.slack.com/) for discussions related to PipeCD development.
- Community Meeting: Every other Wednesdays. Search [here](https://www.cncf.io/calendar/).

### Improvements

- Bug: 
  - Please open an Issue and describe the problem. Or, open a PR with if it's easy one.
- Enhancement / Feature Request:
  - For small changes including docs or adding tests, please open a PR.
  - For new features or big enhancement, basically, please open an Issue and discuss there before sending a PR. We cannot accept all requests in some cases.
- Security issue:
  - Send an email to [the core maintainers](https://github.com/pipe-cd/pipecd/blob/master/SECURITY.md). **DO NOT report on Issues.**

## Release Flow

1. Run the `prepare release` workflow with specifing the new version.
   It will create a PR to update the version.
2. Review the PR and merge it.
3. Create a Release on GitHub with a new tag.
   Then, the `release package` workflow will publish a new npm package.