# cite

> cite is the citation layer for agentic workflows: an open citation graph that humans and AI agents can read, reason over, and contribute back to. Documents are nodes, citations are edges, annotations are how humans and agents build the graph together. This instance exposes a Model Context Protocol (MCP) endpoint so agents can traverse the graph and cite the spans they pulled from. Released as OpenContracts since 2019; rebranded as cite for the v3 release line.

## MCP Server

This instance exposes a read-only MCP server that AI agents can connect to for accessing public corpuses, documents, annotations, and discussion threads.

- Endpoint (global): http://gpt.avalon.click/mcp/
- Endpoint (corpus-scoped): http://gpt.avalon.click/mcp/corpus/{corpus_slug}/
- Protocol: JSON-RPC 2.0 (MCP spec 2025-03-26)
- Auth: None required (public data only)
- Rate limit: 100 requests/minute per IP

### Connecting

Use any MCP-compatible client. For Claude Desktop, add to config:

```json
{
  "mcpServers": {
    "cite": {
      "command": "npx",
      "args": ["mcp-remote", "http://gpt.avalon.click/mcp/"]
    }
  }
}
```

### Available Tools

- `list_public_corpuses`: List all public corpuses (paginated, searchable)
- `list_documents`: List documents in a corpus
- `get_document_text`: Get full extracted text from a document
- `list_annotations`: List annotations on a document (filter by page or label)
- `search_corpus`: Semantic vector search within a corpus
- `list_threads`: List discussion threads in a corpus
- `get_thread_messages`: Get messages in a thread (flat or hierarchical)

### Available Resources (URI-based)

- `corpus://{corpus_slug}` - Corpus metadata
- `document://{corpus_slug}/{document_slug}` - Document with text
- `annotation://{corpus_slug}/{document_slug}/{annotation_id}` - Annotation details
- `thread://{corpus_slug}/threads/{thread_id}` - Discussion thread

## REST Search API

A simple JSON search endpoint at `http://gpt.avalon.click/api/search/?q=QUERY` is available for crawlers and lightweight integrations.

## Links

- [Full MCP documentation](http://gpt.avalon.click/llms-full.txt)
- [Source code](https://github.com/Open-Source-Legal/cite)
- [Project home](https://cite.opensource.legal)