AI Documentation & Guru Knowledge MCP Server

by LaserFocused-ee

249 downloads Not rated yet
GitHub

About

This MCP server provides AI assistants like Claude with structured access to local development documentation and Guru knowledge base content. It serves as a bridge between AI tools and your organization's knowledge repositories, enabling intelligent retrieval of coding guidelines, architecture patterns…

Explore

- Tool-First Architecture: All functionality accessed through well-defined tools
- Dual Knowledge Sources: Access both local docs and Guru knowledge base
- Intelligent Search: Powerful search capabilities across both documentation and Guru
- Rich Metadata: Detailed information about documents and knowledge cards
- Modular Design: Clean separation of concerns for maintainability
- Environment Configuration: Secure credential handling through environment variables
- Error Handling: Graceful degradation with helpful error messages

Setting up with Highlight

This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:

  1. Download and install Highlight from highlightai.com/download
  2. Navigate to the plugins tab and select "Add Custom Plugin"
  3. Configure the plugin with the settings below
    Plugin Name AI Documentation & Guru Knowledge MCP Server
    Command (node, npx, python, etc.)

    Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.

  4. Enable "Start Automatically" if you want the plugin to start when Highlight launches

From the repository

- Node.js: ≥20.0.0
- MCP SDK: v1.12.0 (latest)
- Operating System: Cross-platform (macOS, Linux, Windows)
- Documentation: Markdown files in the expected directory structure
- Guru Access: Valid Guru API credentials (for Guru functionality)

To use Guru tools, configure your environment variables:

{
  "mcpServers": {
    "ai-docs": {
      "command": "node",
      "args": ["dist/index.js"],
      "env": {
        "GURU_TOKEN": "your-username:your-token"
      }
    }
  }
}

Getting your Guru credentials:

1. Log into your Guru account
2. Go to Settings → API Access
3. Generate an API token
4. Format as username:token (e.g., [email protected]:abc123def456)


git clone <repository-url>
cd mcp_server
npm install

bash

The server provides 5 powerful tools for accessing knowledge:

Claude Desktop / Cursor

Paste into your MCP client config file to install this server.

{
    "mcpServers": {
        "ai documentation & guru knowledge mcp server": {
            "ai-docs-mcp": {
                "command": "npx",
                "args": [
                    "@modelcontextprotocol/inspector",
                    "--cli",
                    "node",
                    "dist/index.js",
                    "\\"
                ]
            }
        }
    }
}

McpServers

{
    "ai-docs-mcp": {
        "command": "npx",
        "args": [
            "@modelcontextprotocol/inspector",
            "--cli",
            "node",
            "dist/index.js",
            "\\"
        ]
    }
}

A Model Context Protocol (MCP) server that provides AI assistants like Claude with structured access to both local development documentation and Guru knowledge base content. This server acts as a bridge between AI tools and your organization's knowledge repositories, enabling intelligent retrieval of coding guidelines, architecture patterns, troubleshooting guides, and institutional knowledge.

What is this?

This is an MCP server implementation that:

- Serves structured documentation through the Model Context Protocol standard
- Integrates with Guru knowledge base for accessing team knowledge and procedures
- Provides intelligent document discovery with category-based organization
- Enables AI assistants to access your team's coding standards, runbooks, and troubleshooting guides
- Supports multiple technology stacks including Flutter, React, NestJS, and Git workflows
- Uses modern tool-based architecture for flexible knowledge access

Architecture

The server is built with a clean, modular architecture:

- TypeScript/Node.js runtime (requires Node.js ≥20.0.0)
- MCP SDK v1.12.0 (@modelcontextprotocol/sdk) for protocol implementation
- Modular design with separated concerns for maintainability
- Zod for parameter validation and type safety
- Standard I/O transport for communication with AI clients

Project Structure

src/
├── index.ts              # Main entry point
├── server.ts             # MCP server configuration
├── services/
│   └── guru.ts           # Guru API service
├── tools/
│   ├── docs.ts           # Documentation tools
│   └── guru.ts           # Guru API tools
├── types/
│   └── index.ts          # TypeScript interfaces
└── utils/
    └── docs.ts           # Documentation utilities

Available Tools

The server provides 5 powerful tools for accessing knowledge:

Documentation Tools

legacy-list-docs

Lists all available documentation files with categories

- Returns: Complete inventory of documentation organized by category
- Use case: Discover available documentation before reading specific files

legacy-read-doc

Reads the content of a specific documentation file

- Parameters:
- category (required): Document category (e.g., 'code_guidelines/flutter', 'service-docs')
- name (required): Document name without .md extension
- Returns: Full markdown content of the requested document
- Example: category: "code_guidelines/flutter", name: "best-practices"

Guru Knowledge Base Tools

guru-list-cards

Search and list Guru knowledge cards with advanced filtering

- Parameters:
- searchTerms (optional): Search terms for card title/content
- maxResults (optional): Maximum number of results (max: 50)
- showArchived (optional): Include archived cards
- sortField (optional): Field to sort by (lastModified, title, dateCreated, etc.)
- sortOrder (optional): Sort order (asc/desc)
- query (optional): Advanced Guru Query Language query
- Returns: List of matching cards with metadata (title, collection, owner, status, preview)
- Example: Search for "restarting pods" or "API documentation"

guru-read-card

Read the full content of a specific Guru card

- Parameters:
- cardId (required): The UUID of the Guru card to read
- Returns: Complete card content including HTML formatting, metadata, and attachment URLs
- Use case: Get detailed procedures, troubleshooting steps, or knowledge articles

guru-get-card-attachments

List and optionally download attachments from a Guru card

- Parameters:
- cardId (required): The UUID of the Guru card
- downloadFirst (optional): Whether to download the first attachment content
- Returns: List of attachments with URLs and optional content preview
- Use case: Access screenshots, diagrams, or supplementary files

Example Tool

hello

A simple hello world example demonstrating MCP tool capabilities

- Parameters:
- name (optional): Name to greet
- Returns: Greeting message with usage examples

Environment Configuration

Guru API Setup

To use Guru tools, configure your environment variables:

{
  "mcpServers": {
    "ai-docs": {
      "command": "node",
      "args": ["dist/index.js"],
      "env": {
        "GURU_TOKEN": "your-username:your-token"
      }
    }
  }
}

Getting your Guru credentials:

1. Log into your Guru account
2. Go to Settings → API Access
3. Generate an API token
4. Format as username:token (e.g., [email protected]:abc123def456)

Documentation Structure

git stdsdThe server includes documentation in this structure:

docs/
├── code_guidelines/
│   ├── flutter/
│   │   ├── architecture/
│   │   │   ├── flutter_best_practices.md
│   │   │   ├── flutter_riverpod_clean_architecture.md
│   │   │   └── freezed_guide.md
│   │   └── riverpod_testing_guide.md
│   ├── react/
│   │   └── react_best_practices.md
│   ├── nestjs/
│   │   └── nestjs_best_practices.md
│   └── git/
│       └── graphite_commands_reference.md
└── service-docs/
    ├── claude-code-sdk-documentation.md
    ├── guru-api-integration-guide.md
    ├── linear-sdk-documentation.md
    ├── mcp-inspector-testing-guide.md
    └── mcp-server-scaffold-guide.md

Installation & Usage

Development Setup

```bash

No reviews yet — be the first

Sign in to leave a review

Use Google, GitHub, or an email account so ratings stay tied to real people.

Email sign in

No reviews posted yet.