Skip to main content
Glama

@edziocodes/niivue-mcp

npm version

MCP server that provides Niivue documentation and API reference to LLMs with fast lexical search and optional semantic hybrid search.

What you get

  • Full text search over Niivue docs and guides (BM25 + embeddings).

  • API lookup from JSDoc comments in the Niivue TypeScript source.

  • Structured tools for overview, search, listing, and page retrieval.

  • Local cache to keep startup and query latency low.

Related MCP server: qkb

Installation

npm install @edziocodes/niivue-mcp

MCP Client Configuration

Add the following to your MCP client configuration (e.g., Claude Desktop, Cursor, etc.):

{
  "mcpServers": {
    "niivue": {
      "command": "npx",
      "args": ["-y", "@edziocodes/niivue-mcp"]
    }
  }
}

Development

Clone and build from source:

git clone https://github.com/gaiborjosue/niivue-mcp.git
cd niivue-mcp
npm install
npm run build

Run locally:

npm run dev

Hybrid search with semantic embeddings is enabled by default. The embedding model is downloaded on first use.

To disable embeddings and use BM25 search only:

npx @edziocodes/niivue-mcp --no-embeddings

Tools

  • get_niivue_overview

  • search_niivue_docs

  • get_niivue_doc

  • list_niivue_docs

  • search_niivue_api

  • get_niivue_api

Cache

Stored in ~/.niivue-mcp:

  • docs/ raw markdown

  • index.json BM25 index

  • vectors.json vector index (when embeddings enabled)

  • api-index.json parsed API entries

  • meta.json cache metadata

Attribution

Docs and source are fetched from the Niivue repository: https://github.com/niivue/niivue/

Available Tools

6 tools
get_niivue_apiC

Fetch full details for a specific Niivue API method.

Inputs: name (required).

get_niivue_docC

Fetch full content for a Niivue doc page.

Inputs: page (required).

get_niivue_overviewB

Return the CLAUDE.md overview for Niivue.

Inputs: no named parameters.

list_niivue_docsB

List available Niivue docs and metadata.

Inputs: no named parameters.

search_niivue_apiC

Search Niivue API methods by name or description.

Inputs: query (required), limit.

search_niivue_docsC

Search Niivue documentation with BM25.

Inputs: query (required), limit.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A3.5/5.0

    Scored across 6 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose with no ambiguity: get_niivue_api fetches API method details, get_niivue_doc fetches doc page content, get_niivue_overview returns the overview, list_niivue_docs lists docs, search_niivue_api searches API methods, and search_niivue_docs searches documentation. The tools are well-differentiated by their target resource (API vs. docs) and action (get vs. list vs. search).

    Naming Consistency5/5

    All tool names follow a consistent verb_noun pattern with 'niivue' as a prefix, using snake_case throughout. The verbs (get, list, search) are applied predictably to the nouns (api, doc, overview, docs), making the naming highly consistent and readable.

    Tool Count5/5

    With 6 tools, the server is well-scoped for its purpose of accessing Niivue documentation and API information. Each tool earns its place by covering distinct aspects like fetching, listing, and searching, without being too sparse or bloated, which is ideal for a documentation-focused server.

    Completeness4/5

    The tool set provides comprehensive coverage for accessing and searching Niivue documentation and API details, with no obvious gaps for the stated purpose. Minor gaps might include lack of update or delete operations, but these are not expected in a read-only documentation context, so the surface is nearly complete.

    Maintenance

    ActivityInactive
    ResponsivenessNo recent community issues

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      B
      maintenance
      Provides a hybrid search engine for Obsidian vaults, enabling LLM agents to query notes with BM25 keyword and vector semantic search, metadata filtering, and sibling-document retrieval.
      265 npm
      MIT
    • A
      license
      A
      quality
      C
      maintenance
      Enables AI coding agents to fetch live MDN Web Docs pages, TypeScript handbook content, and npm package metadata as clean markdown, with local search, caching, and polite throttling.
      5
      MIT