MemTensor/memmy-agent

What Is Memmy?

Cross-Agent Task Continuity

Most of the Agents You're Using Can Connect to Memmy

DeepSeek Harness, OpenClaw, Hermes, Claude Code, Codex, Cursor, WorkBuddy, OpenCode, Pi...they all work!

Data Security

How to Use Memmy

For complete installation and configuration instructions, see the Getting Started guide.

1. Desktop App (Recommended)

Download Memmy from the official website or GitHub Releases.

[!TIP]

Sign up for Memmy to receive free tokens and try the complete Memory + Agent Runtime.

Trial credits:

Registration grants Agent task trial tokens; the current balance and usage are shown in the app.

When the trial credits run out, switch to BYOK mode and use your own model API.

2. Use the memmy CLI / TUI

On Linux x64 or arm64 with Node.js 22 or newer and an available systemd user session:

curl -fsSL https://raw.githubusercontent.com/MemTensor/memmy-agent/main/scripts/install.sh | bash
memmy

The installer enables the local Memory Service immediately as memmy-memory.service. The first bare memmy invocation opens the model setup wizard when needed, then enables memmy-gateway.service, waits for it to become ready, and enters the TUI. Both are systemd --user services bound to localhost and remain available after the TUI or terminal exits. They start again on later logins; the installer does not enable linger. Only the installer launcher activates this service management, so source-built Linux CLIs keep their existing behavior.

Before starting or reconnecting to the Gateway, memmy refreshes a private ~/.memmy/systemd/gateway.env file (mode 0600) with configuration-referenced environment variables, common Provider credentials, and the terminal PATH. If those values change, the next bare memmy invocation restarts the user service with the new environment.

systemctl --user status memmy-memory.service
systemctl --user status memmy-gateway.service

The installer initializes Memory without changing Codex, Claude Code, Cursor, or other agents. Run memmy-memory init (all detected agents) or memmy-memory init --agent when you explicitly want to install the Memory Skill and the supported Hook/plugin for an agent.

memmy onboard                              # Configure models, providers, gateway, memory, and tools interactively
memmy onboard --defaults                   # Initialize ~/.memmy/config.yaml and the workspace with defaults
memmy status                               # Check the configuration, model, and provider
memmy agent --message "Introduce the current workspace"  # Run a single-turn task
memmy                                      # Enter the interactive TUI
memmy serve                                # Start the OpenAI-compatible API (:18990)

The minimal BYOK configuration is located at ~/.memmy/config.yaml:

agents:
  defaults:
    model: openai/gpt-4.1
    provider: openai
    timezone: "+08:00"
providers:
  openai:
    apiKey: ${OPENAI_API_KEY}

3. Use the memmy-memory CLI

Use it to access the local memory service from agents, scripts, and debugging workflows:

memmy-memory init
memmy-memory health
memmy-memory search "memory policies in this project"
memmy-memory add "a piece of knowledge worth saving"
memmy-memory get 

It connects to http://127.0.0.1:18960 by default. Use --url, --token, --config, --source, and --user-id to specify the service and namespace.

4. Start from the Source Code

git clone https://github.com/MemTensor/memmy-agent.git
cd memmy-agent
cp .env.example .env
npm install
npm run build
bash scripts/dev-start.sh

The script installs dependencies, builds the services, and starts the development environment. Node.js >=22 and npm are required; use Git Bash on Windows.

How Is Memmy Built?

For details about the architecture, memory service, and integration methods, see the Memmy documentation.

Roadmap

Memmy is building personal memory infrastructure, and its scope goes beyond coding Agents:

• More memory sources — expanding from AI conversations to browser activity, local documents, and eventually more devices and hardware.
• Team collaboration — planned Agent-to-Agent collaboration, letting team members' AI assistants share knowledge under privacy protection.

Acknowledgements

Memmy stands on the shoulders of a group of excellent open-source projects, and we are deeply grateful.

• OpenClaw — a pioneer of open-source personal AI assistants; its exploration of multi-platform messaging channels directly inspired Memmy's channel connection design.
• hermes-agent — the self-evolving Agent built by Nous Research; its practice in persistent memory and skill self-learning showed us that an Agent can "understand you better the more you use it".
• nanobot — grown from a minimal prototype into a fully featured open-source Agent platform; its engineering practice around the Agent loop and MCP integration provided important references for Memmy's core design.

The point of open source is to let good ideas flow, and we hope Memmy becomes part of that river.

Contributors

Thanks to every contributor who makes Memmy better ❤️