kentcdodds/kody

Your assistant's home — the memory, keys, code, and automations your AI agent keeps, portable across every MCP host. Built on Cloudflare Workers.

What it solves

Kody provides a portable "home" for AI agents, allowing them to maintain persistent memory, store API keys, manage code, and run automations across different Model Context Protocol (MCP) hosts. It solves the problem of agent state being locked into a single platform by making the assistant's data and tools portable.

How it works

Built on Cloudflare Workers and the Model Context Protocol (MCP), Kody acts as a server that AI agents connect to. It uses a Remix UI for management and Cloudflare D1, KV, and Durable Objects to handle data storage and state. The system prioritizes a compact set of powerful tools—specifically search and code execution—over a large list of static tools.

Who it’s for

Users who want a multi-user personal assistant platform where each user has an isolated environment (including secrets and memories) that can be accessed by various MCP-compatible AI hosts, such as ChatGPT.

Highlights

  • MCP-First Architecture: Designed to be portable across any compatible MCP host.
  • Isolated User State: Each signed-in user gets their own dedicated packages, jobs, secrets, and memories.
  • Cloudflare Native: Fully leverages Cloudflare Workers, D1 database, and Durable Objects for scale and performance.
  • Code Mode Execution: Features a specialized flow for executing code as part of the agent's capabilities.