# graphlit-mcp-server [Health: Active]

**Category:** 🧠 Knowledge & Memory  
**Repository:** https://github.com/graphlit/graphlit-mcp-server  
**GitHub Stars:** 379  
**npm Downloads (last month):** 800  
**Views:** 4  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/graphlit-mcp-server

## Description
Ingest anything from Slack, Discord, websites, Google Drive, Linear or GitHub into a Graphlit project - and then search and retrieve relevant knowledge within an MCP client like Cursor, Windsurf or Cline.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "graphlit-mcp-server": {
    "command": "npx",
    "args": ["-y","graphlit-mcp-server"],
    "env": {
      "GRAPHLIT_ENVIRONMENT_ID": "",
      "GRAPHLIT_ORGANIZATION_ID": "",
      "GRAPHLIT_JWT_SECRET": ""
    }
  }
}
```

**Requires environment variables:** `GRAPHLIT_ENVIRONMENT_ID`, `GRAPHLIT_ORGANIZATION_ID`, `GRAPHLIT_JWT_SECRET` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What graphlit-mcp-server MCP server does

The graphlit-mcp-server MCP server exposes Graphlit capabilities through the Model Context Protocol. It lets an MCP client add content to a Graphlit project, search that project, retrieve relevant sources, and manage collections, feeds, conversations, and other project resources.

Supported ingestion types include files, web pages, messages, posts, emails, issues, text, and short-term memory. The listed connectors cover Slack, Discord, Microsoft Teams, Outlook, Gmail, Notion, Reddit, Linear, Jira, GitHub, Google Drive, OneDrive, SharePoint, Dropbox, Box, Twitter/X, and podcast RSS feeds. The material also identifies built-in web crawling, web search, web mapping, and page screenshots.

## How it works

The server uses a Graphlit project as the shared knowledge store. Content imported into that project can be queried through MCP tools, including searches across contents, collections, feeds, and conversations. Agents can also retrieve relevant sources, find similar images, visually describe images, prompt an LLM conversation, and extract structured JSON from text.

During ingestion, PDF, DOCX, PPTX, and other documents are extracted into Markdown. HTML pages are handled similarly, while audio and video are transcribed. Graphlit operations also include creating and editing collections, deleting content or feeds, checking processing status, and configuring the project.

## Setup and configuration

The graphlit-mcp-server MCP server requires Node.js, with version 18 or later recommended, and an active Graphlit Platform account with access to the API settings dashboard. Install the npm package through an MCP client configuration using `npx -y graphlit-mcp-server`.

Set these environment variables for authentication and project configuration:

- `GRAPHLIT_ENVIRONMENT_ID`: Graphlit environment identifier.
- `GRAPHLIT_ORGANIZATION_ID`: Graphlit organization identifier.
- `GRAPHLIT_JWT_SECRET`: Secret used to sign the JWT token.

The README provides configurations for VS Code, Windsurf, Cline, and Cursor. Values for the organization ID, environment ID, and JWT secret are obtained from the Graphlit Platform API settings dashboard.

## Tools and capabilities

Available tool groups include retrieval, RAG, extraction, publishing, ingestion, data connectors, web operations, notifications, project operations, and enumerations. Publishing tools can produce audio through ElevenLabs or images through OpenAI image generation. Notification destinations include Slack, email, webhooks, and Twitter/X.

Enumeration tools can list resources such as Slack channels, Microsoft Teams teams and channels, SharePoint libraries and folders, Linear projects, Notion databases and pages, cloud-storage folders, Discord guilds and channels, and Google or Microsoft calendars. The server also exposes MCP resources for projects, contents, feeds, collections, workflows, conversations, and specifications.

## Limitations and notes

The graphlit-mcp-server MCP server depends on the Graphlit Platform account and its configured credentials; the provided material does not describe an offline or local-only mode. Connector availability may depend on the corresponding Graphlit project configuration, but the README does not specify separate setup steps for each source. Claude Desktop compatibility and pricing are not stated in the provided material.

_Full upstream README: https://allmcps.com/mcp/graphlit-mcp-server/readme_