# Engram

> Engram is an MCP server for knowledge and memory. Persistent memory for coding agents across sessions, repos and teammates. It can run locally on your machine or as a hosted remote server.

- **Kind:** MCP server
- **Pricing:** Free trial
- **Platforms:** CLI, API
- **Website:** https://engram.tools/en/agent
- **MCP Registry name:** tools.engram/engram
- **Categories:** knowledge and memory, developer tools
- **Listing updated:** 2026-11-16

## About

Engram is built around a memory loop the agent is instructed to follow. A session opens with `engram_start`, passing project, repo, role, stack and task, which loads setup guidance, relevant memories and matching skill guides. Before choosing an approach the agent calls `engram_recall`; if another repo may have solved the problem it uses `engram_cross_pollinate`; after context compaction it calls `engram_rehydrate`; and at milestones `engram_reflect` saves a checkpoint and reports which recalled memories helped.

Memories have types such as solution or convention, carry tags and repo context, and can be withdrawn reversibly and restored after revalidation. Reporting whether a memory helped, was stale or came from the wrong context trains future ranking. The site's examples include a frontend agent in Cursor passing an API convention to a backend agent in Claude Code.

The npm package starts an isolated free trial with no email or key, and running its connect command later claims the workspace and keeps memories. Remote HTTP connections use OAuth with email verification, and a human approves paid upgrades. The site says it exposes 17 tools and advises keeping secrets out of memory.

## Features

- **Session start:** One call loads guidance, relevant memories and skill guides.
- **Semantic recall:** Retrieve memories filtered by repo, type or tags.
- **Cross-repo learning:** Surface lessons from other repositories in your workspace.
- **Context rehydration:** Recover the working map after compaction or resume.
- **Outcome feedback:** Report whether a memory helped to improve ranking.
- **Reversible retraction:** Withdraw stale advice and restore it after revalidation.

## Good for

- Remember a project decision and why across sessions
- Share a convention discovered by one agent with another
- Recover context after a long session is compacted
- Apply a fix from one repo to a similar problem in another
- Hand off work between teammates' agents

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http engram https://api.engram.tools/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://api.engram.tools/mcp
```

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

```json
{
  "mcpServers": {
    "engram": {
      "url": "https://api.engram.tools/mcp"
    }
  }
}
```

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

```json
{
  "servers": {
    "engram": {
      "type": "http",
      "url": "https://api.engram.tools/mcp"
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "engram": {
      "serverUrl": "https://api.engram.tools/mcp"
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "engram": {
      "type": "http",
      "url": "https://api.engram.tools/mcp"
    }
  }
}
```

## Tools

- `engram_start`: Start a session and load context
- `engram_recall`: Recall relevant memories by semantic query
- `engram_cross_pollinate`: Find lessons from other repos
- `engram_rehydrate`: Recover context after compaction
- `engram_reflect`: Save a milestone checkpoint and learnings
- `engram_store`: Store a convention, pattern, gotcha or solution
- `engram_report_outcome`: Report whether a recalled memory helped
- `engram_proof`: Return product proof and usage figures

## FAQ

### Can I try it without an email?

Yes, the npm package starts an isolated free trial with no email or API key.

### Which clients are supported?

Claude, Codex, Cursor, Gemini or Antigravity, Windsurf and VS Code.

### How do I keep the trial's memories?

Run engram-tools connect later and verify your email.

### Who approves paid upgrades?

A human does.

### Should secrets be stored?

No, the instructions say to keep secrets out of memory.

### Is Engram free?

Engram is a paid product with a free trial.

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

Run this command in your terminal: claude mcp add --transport http engram https://api.engram.tools/mcp

### What are the best alternatives to Engram?

The closest alternatives to Engram are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [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.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Ouroboros](https://ai-powered.com/mcp/ouroboros): Ouroboros is an open-source MCP server for developer tools.
- [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.

---
Source: https://ai-powered.com/mcp/engram-10
