# Memnest

> Memnest is a free MCP server for knowledge and memory. Local memory for coding AIs that searches your original conversations, not summaries. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Free
- **Platforms:** macOS, Linux, Windows, CLI
- **Repository:** https://github.com/blue-b/memnest
- **MCP Registry name:** io.github.Blue-B/memnest
- **Categories:** [knowledge and memory](https://ai-powered.com/categories/knowledge-and-memory), [developer tools](https://ai-powered.com/categories/developer-tools)
- **Listing updated:** 2026-11-15

## About

Memnest stores memories exactly as written and keeps them on your machine in a single Rust binary with SQLite and a local multilingual embedding model. Saving and searching call no model API, so there is no key and no extra token bill. Search merges keyword and vector rankings, favours memories saved on purpose over raw turns, and returns three results by default, each with a source ID.

One local service is shared by Claude Code, Codex, Cursor and pi, so a decision made in one is findable in another. A conversation watcher can capture visible turns from Claude Code, Codex and pi, and an encrypted vault holds credentials out of search. Task cards carry multi-day work between sessions. Automatic recall is off unless you turn it on.

The first search downloads an embedding model of about 1.1 GB. It supports macOS on Apple Silicon, Linux with systemd and 64-bit Windows 10 or 11. The README notes search can miss records and cannot tell whether an old decision still holds.

## Features

- **Original-text search:** Keyword and vector search over stored text, with source IDs.
- **Shared across AIs:** Claude Code, Codex, Cursor and pi use one local memory.
- **Task cards:** Carry multi-day work between sessions.
- **Encrypted vault:** Keeps credentials out of search results.
- **Local-only:** Runs on your machine with no model API calls.

## Good for

- Recall why a past decision was made
- Share project rules across several coding agents
- Resume a multi-day migration next session
- Search all projects for an earlier decision

## Install

### Claude Code (Terminal)

```sh
claude mcp add memnest -- npx -y https://github.com/Blue-B/memnest/releases/download/v0.6.0/memnest-v0.6.0-linux-x64.mcpb
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "memnest": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Blue-B/memnest/releases/download/v0.6.0/memnest-v0.6.0-linux-x64.mcpb"
      ]
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "memnest": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Blue-B/memnest/releases/download/v0.6.0/memnest-v0.6.0-linux-x64.mcpb"
      ]
    }
  }
}
```

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

```json
{
  "servers": {
    "memnest": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Blue-B/memnest/releases/download/v0.6.0/memnest-v0.6.0-linux-x64.mcpb"
      ]
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "memnest": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Blue-B/memnest/releases/download/v0.6.0/memnest-v0.6.0-linux-x64.mcpb"
      ]
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "memnest": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/Blue-B/memnest/releases/download/v0.6.0/memnest-v0.6.0-linux-x64.mcpb"
      ]
    }
  }
}
```

## Tools

- `memory_search`: Searches memories and returns stored text with IDs.
- `memory_get`: Opens the original record.
- `memory_remember`: Saves a memory as written.
- `memory_update`: Replaces a memory and keeps history.
- `memory_delete`: Deletes a memory.
- `task_card`: Saves and reopens a plan for multi-day work.

## FAQ

### Does it need an API key?

No. Saving and searching call no model API.

### What does it download?

An embedding model of about 1.1 GB on first use.

### Which platforms are supported?

macOS on Apple Silicon, Linux with systemd and 64-bit Windows 10 or 11.

### Is automatic recall on by default?

No, it is off unless you enable it.

### Is Memnest free?

Yes. Memnest is free to use.

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

Run this command in your terminal: claude mcp add memnest -- npx -y https://github.com/Blue-B/memnest/releases/download/v0.6.0/memnest-v0.6.0-linux-x64.mcpb

### What are the best alternatives to Memnest?

The closest alternatives to Memnest are REA, Context7, Chrome DevTools MCP, Codebase Memory and GitHub.

## Where this comes from

- Description: written from the project's README.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Context7](https://ai-powered.com/mcp/context7): Context7 is a freemium MCP server for developer tools.
- [Chrome DevTools MCP](https://ai-powered.com/mcp/chrome-devtools-mcp-2): Chrome DevTools MCP is an open-source MCP server for browser automation.
- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [GitHub](https://ai-powered.com/mcp/github-2): GitHub is an open-source MCP server for developer tools.
- [Serena MCP: the IDE for your agent](https://ai-powered.com/mcp/serena-mcp-the-ide-for-your-agent): Serena MCP: the IDE for your agent is an open-source MCP server for developer tools.
- [awesome-mcp-servers](https://ai-powered.com/mcp/awesome-mcp-servers-2): awesome-mcp-servers is an open-source MCP server for developer tools.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.

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