Claude Recall
A persistent memory server for Claude Code that captures session context and tool outputs to inject relevant history into future sessions. It enables long-term recall through semantic search and automatic context management, allowing for more consist
What is Claude Recall?
Claude Recall is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to persistent memory server for claude code that captures session context and tool outputs to inject relevant history into future sessions. it enables long-term recall through semantic search and automat...
A persistent memory server for Claude Code that captures session context and tool outputs to inject relevant history into future sessions. It enables long-term recall through semantic search and automatic context management, allowing for more consist
This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- A persistent memory server for Claude Code that captures ses
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx claude-recallConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Claude Recall
Claude Recall (MoltBrain) is a persistent memory server for Claude Code that automatically captures session observations — including tool outputs, design decisions, and discovered code patterns — and stores them with semantic vector embeddings so they can be retrieved in future sessions. It uses ChromaDB for vector storage and exposes a local web viewer and REST API at localhost:37777, giving developers a searchable timeline of everything Claude has learned across projects. By injecting relevant past context into new sessions automatically, it enables more consistent long-running workflows where Claude remembers architectural decisions and project-specific conventions.
Prerequisites
- Node.js 18 or higher, or Bun 1.0 or higher
- Claude Code installed and configured
- ChromaDB accessible (bundled or separately installed) for vector embeddings
- MoltBrain plugin installed via the Claude Code plugin marketplace
- Optional: ANTHROPIC_API_KEY if using Claude as the summary model
Install MoltBrain via the Claude Code plugin marketplace
Add the MoltBrain plugin directly from the Claude Code plugin marketplace. The system activates automatically with no additional configuration required after installation.
/plugin marketplace add nhevers/moltbrain
/plugin install moltbrainVerify the worker service is running
After installation, the MoltBrain worker service should start automatically on port 37777. Check the health endpoint to confirm it is active.
curl http://localhost:37777/healthConfigure settings (optional)
Adjust memory behavior by editing ~/.moltbrain/settings.json. Key settings include the maximum number of memory injections per session and the AI model used for generating summaries.
# ~/.moltbrain/settings.json
{
"MOLTBRAIN_WORKER_PORT": 37777,
"MOLTBRAIN_CONTEXT_OBSERVATIONS": 50,
"MOLTBRAIN_PROVIDER": "claude",
"MOLTBRAIN_PRUNE_DAYS": 0,
"MOLTBRAIN_LOG_LEVEL": "info"
}Search stored memories from the CLI
Use the moltbrain CLI to manually search for past observations using semantic search against stored vector embeddings.
moltbrain search "authentication middleware"
moltbrain stats
moltbrain exportOrganize observations with tags
Tag specific observations to make them easier to find and filter later. Use the observation ID shown in the web viewer or CLI output.
moltbrain tag <observation-id> architecture
moltbrain tag <observation-id> bug-fixOpen the web viewer
Browse your full memory timeline, analytics, and session history using the built-in web interface.
open http://localhost:37777Claude Recall Examples
Client configuration
MCP server configuration for Claude Desktop to connect to the MoltBrain memory worker. The worker must be running locally first.
{
"mcpServers": {
"claude-recall": {
"command": "npx",
"args": ["claude-recall"],
"env": {
"MOLTBRAIN_WORKER_PORT": "37777",
"MOLTBRAIN_CONTEXT_OBSERVATIONS": "50",
"MOLTBRAIN_PROVIDER": "claude"
}
}
}
}Prompts to try
Example prompts that leverage persistent memory in Claude Code sessions
- "What did we decide about the database schema in previous sessions?"
- "Search memory for any notes about the authentication flow"
- "Summarize what you remember about this project's architecture"
- "Tag this observation as a key architectural decision"Troubleshooting Claude Recall
Worker service not starting or health check fails
Check if port 37777 is already in use with lsof -i :37777. Change MOLTBRAIN_WORKER_PORT in ~/.moltbrain/settings.json if there is a conflict. Also confirm Node.js 18+ is installed with node --version.
Memory observations are not being injected into new sessions
Verify MOLTBRAIN_CONTEXT_OBSERVATIONS is not set to 0. Check that ChromaDB is running and accessible. Run moltbrain stats to confirm observations have been stored from previous sessions.
Semantic search returns irrelevant results
The embedding model improves with more stored observations. Use moltbrain tag to label high-quality observations, and use moltbrain prune to remove low-quality or outdated ones. Try more specific search queries.
Frequently Asked Questions about Claude Recall
What is Claude Recall?
Claude Recall is a Model Context Protocol (MCP) server that persistent memory server for claude code that captures session context and tool outputs to inject relevant history into future sessions. it enables long-term recall through semantic search and automatic context management, allowing for more consist It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Claude Recall?
Follow the installation instructions on the Claude Recall GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Claude Recall?
Claude Recall works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Claude Recall free to use?
Yes, Claude Recall is open source and available under the ISC License license. You can use it freely in both personal and commercial projects.
Claude Recall Alternatives — Similar Knowledge & Memory Servers
Looking for alternatives to Claude Recall? Here are other popular knowledge & memory servers you can use with Claude, Cursor, and VS Code.
MemPalace
★ 52.6kA local AI memory system that stores all conversations verbatim and organizes them into navigable structures. It provides 19 MCP tools for AI assistants to search and retrieve past decisions, debugging sessions, and architecture debates automatically
Kratos
★ 25.7k🏛️ Memory System for AI Coding Tools - Never explain your codebase again. MCP server with perfect project isolation, 95.8% context accuracy, and the Four Pillars Framework.
Context Mode
★ 15.4kAn MCP server that preserves LLM context by intercepting large data outputs and returning only concise summaries or relevant sections. It enables efficient sandboxed code execution, file processing, and documentation indexing across multiple programm
Memu
★ 13.7kMemory for 24/7 proactive agents like OpenClaw.
MemOS
★ 9.3kMemOS (Memory Operating System) is a memory management operating system designed for AI applications. Its goal is: to enable your AI system to have long-term memory like a human, not only remembering what users have said but also actively invoking, u
Everos
★ 5.4kBuild, evaluate, and integrate long-term memory for self-evolving agents.
Browse More Knowledge & Memory MCP Servers
Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Claude Recall in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Claude Recall?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.