Desjajja/docuscribe-mcp-server

Docuscribe MCP

Overview

Docuscribe MCP is a tool designed to interact with documentation libraries using MCP (Model Context Protocol). It provides two main tools for querying and fetching content:

1. list_all_docs: Retrieves a pageable catalog of all available documentation libraries with stable IDs.
2. fetch_doc_content: Fetches content from a specific document in three modes: index, single range, and multi-range.

Tools

list_all_docs

Description

Retrieves a pageable catalog of all available documentation libraries (documents) with stable IDs (uids).

Parameters

• limit (int, optional): Max number of documents (1..1000, default 100).
• offset (int, optional): Pagination offset (>=0, default 0).

Returns

A JSON string exactly mirroring backend response:

{
  "documents": [
    {"id": "uuid", "name": "Title", "hashtags": ["tag1", "tag2"]},
    ...
  ]
}

Usage Guidance

• Call this first to discover a document's stable id.
• Use that id as doc_uid for fetch_doc_content.
• Persist (cache) returned ids; they do not change during session.
• Avoid re-calling once ids are known unless searching for new libraries or pagination needed.
• May store multiple ids for multi-source comparison.

Examples

• Minimal call:

  list_all_docs(limit=50, offset=0)
  

• Paginated follow-up:

  list_all_docs(limit=100, offset=100)
  

fetch_doc_content

Description

Fetches documentation content in three modes:

1. Index Mode: No parameters -> structural overview (pages, ranges).
2. Single Range Mode: start + max_length -> sequential slice with paging hints (has_more, next_start).
3. Multi-Range Mode: ranges parameter -> targeted non-contiguous retrieval.

Parameters

• doc_uid (str): Stable document ID from list_all_docs.
• start (int, optional): Starting word index (Single Range Mode).
• max_length (int, optional): Max words to return (Single Range).
• ranges (str, optional): Comma-separated list of start-end pairs (e.g., "0-100,200-300") for Multi-Range Mode.

Returns

A JSON string combining backend document and meta. Convenience keys has_more and next_start are surfaced when present.

Usage Guidance

• Step 2: After caching doc_uid, call Index Mode.
• Analyze pages/ranges to identify candidate section(s).
• If uncertain: begin Single Range at start=0; iterate while has_more and info not found.
• Use keyword/heading detection to decide early switch to targeted Multi-Range.
• Accumulate content locally; avoid refetching previously retrieved spans.

Examples

• Index retrieval:

  fetch_doc_content(doc_uid="abc123")
  

• Progressive Single Range loop:

  fetch_doc_content(doc_uid="abc123", start=0, max_length=500)
  

• Multi-Range targeted fetch:

  fetch_doc_content(doc_uid="abc123", ranges="0-120,400-550")
  

Installation

1. Clone the repository:

   git clone <repository-url>
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Set up environment variables:
   • Create a .env.local file with the following content:

     SERVER_BACKEND_URL=http://localhost:9002
     

Running the MCP Server

Start the MCP server:

python main.py

Real Use Cases

Notes

• Ensure the backend server is running at the URL specified in .env.local.
• Use list_all_docs to retrieve document IDs before calling fetch_doc_content.
• Cache document IDs to optimize queries and avoid unnecessary lookups.

License

MIT License