# MaxGraph

> MaxGraph is an MCP server for knowledge and memory. Self-hosted long-term memory for AI agents with a knowledge graph and hybrid search. It runs as a hosted remote server, so there is nothing to install locally.

- **Kind:** MCP server
- **Pricing:** Free trial
- **Platforms:** Windows, macOS, Linux, self-hosted
- **Website:** https://rezon-ai.ru
- **Repository:** https://github.com/rumemory/maxgraph
- **MCP Registry name:** io.github.RuMemory/maxgraph
- **Categories:** knowledge and memory, AI models and LLM tooling
- **Listing updated:** 2026-11-15

## About

MaxGraph keeps project memory outside the chat. Decisions, bugs, facts and open tasks are written to plain files on your own computer as they come up, so they survive closed or compacted conversations. Each new session starts with a short summary of past decisions, open tasks and previous session results, and a live dashboard shows a map of connections, a feed and search.

It is aimed at project work rather than personal preferences: the README contrasts it with ChatGPT and Claude memory, which remember you, and says MaxGraph remembers your projects. Memory is split by project so work code, personal notes and freelance work stay separate. Search finds records by meaning, and all data can be exported as NDJSON. The site claims about 43 percent fewer tokens used, since history is not repeated in each message.

It connects to Claude Code through a plugin, to Cursor and VS Code by one-click links, and to Cline, Continue and other clients over streamable HTTP with a bearer token. REST works for ChatGPT-compatible and other models. Install runs on Linux, macOS and Windows without Docker, and an OCI image is also listed. The site is in Russian and offers a 7-day trial with no card.

## Features

- **Automatic memory capture:** Agents record decisions, bugs, facts and tasks as they appear.
- **Session start summary:** New chats begin with past decisions and open tasks.
- **Semantic search:** Finds records by meaning, not exact words.
- **Live dashboard:** A map of connections, a dated feed and search over every record.
- **Project separation:** Memory is divided into projects.
- **NDJSON export:** Export all data so nothing is locked in.

## Good for

- Resume a coding project in a new chat without re-explaining
- Share project context across Claude, Cursor and other agents
- Track open bugs and tasks across sessions
- Audit what an agent has remembered

## Install

### Claude Code (Terminal)

```sh
claude mcp add maxgraph --env MEMORY_TOKEN='<MEMORY_TOKEN>' -- docker run -i --rm -e MEMORY_TOKEN ghcr.io/rumemory/maxgraph:1.9.5
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "maxgraph": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "MEMORY_TOKEN",
        "ghcr.io/rumemory/maxgraph:1.9.5"
      ],
      "env": {
        "MEMORY_TOKEN": "<MEMORY_TOKEN>"
      }
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "maxgraph": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "MEMORY_TOKEN",
        "ghcr.io/rumemory/maxgraph:1.9.5"
      ],
      "env": {
        "MEMORY_TOKEN": "<MEMORY_TOKEN>"
      }
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "maxgraph": {
      "type": "stdio",
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "MEMORY_TOKEN",
        "ghcr.io/rumemory/maxgraph:1.9.5"
      ],
      "env": {
        "MEMORY_TOKEN": "<MEMORY_TOKEN>"
      }
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "maxgraph": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "MEMORY_TOKEN",
        "ghcr.io/rumemory/maxgraph:1.9.5"
      ],
      "env": {
        "MEMORY_TOKEN": "<MEMORY_TOKEN>"
      }
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "maxgraph": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "MEMORY_TOKEN",
        "ghcr.io/rumemory/maxgraph:1.9.5"
      ],
      "env": {
        "MEMORY_TOKEN": "<MEMORY_TOKEN>"
      }
    }
  }
}
```

## FAQ

### Where is the data stored?

As plain files on your own computer, according to the README.

### Which clients can use it?

Claude Desktop and Code, Cursor, VS Code, Cline, Continue, plus REST clients for ChatGPT and other models.

### Is there a trial?

Yes, 7 days with no card, starting on first launch.

### Does it need Docker?

No. The README gives a one-command installer for Linux, macOS and Windows.

### Can I take my data out?

Yes, the site says all data exports to NDJSON with one button.

### Is MaxGraph free?

MaxGraph is a paid product with a free trial.

### How do I add MaxGraph to Claude Code?

Run this command in your terminal: claude mcp add maxgraph --env MEMORY_TOKEN='<MEMORY_TOKEN>' -- docker run -i --rm -e MEMORY_TOKEN ghcr.io/rumemory/maxgraph:1.9.5

### What are the best alternatives to MaxGraph?

The closest alternatives to MaxGraph are Codebase Memory, XERJ, Claude Code Ultimate Guide, Nautilus Compass Memory and note1.

## Alternatives

- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Nautilus Compass Memory](https://ai-powered.com/mcp/nautilus-compass-memory): Nautilus Compass Memory is an open-source MCP server for knowledge and memory.
- [note1](https://ai-powered.com/mcp/note1): note1 is an MCP server for productivity.
- [MedXpert 图书馆连接器（通用版）](https://ai-powered.com/mcp/medxpert): MedXpert 图书馆连接器（通用版） is an open-source MCP server for knowledge and memory.
- [Context Nest](https://ai-powered.com/mcp/context-nest): Context Nest is a freemium MCP server for knowledge and memory.
- [Tulsa Ecosystem Map](https://ai-powered.com/mcp/tulsa-ecosystem-map): Tulsa Ecosystem Map is a free MCP server for knowledge and memory.

---
Source: https://ai-powered.com/mcp/maxgraph
