ENZO: An open-source, self-hosted AI platform for interacting with 300+ models

Introduction to ENZO

ENZO is an open-source, self-hosted AI platform designed for local use. It enables users to interact with a wide range of AI models, specifically over 300, and develop AI agents that can generate their own operating manuals. The platform focuses on providing a direct interface for AI interactions without third-party intervention.

Key Features and Operation

The platform allows users to conduct research, generate code, and execute it within their own infrastructure. When a user sends a message, the request travels from their browser through ENZO directly to the selected AI provider. Users pay the provider's standard price, as ENZO does not take a cut, nor does it require an account, usage meter, or subscription.

Installation is simplified, requiring a `git clone` and `docker compose up -d` to run locally. Provider keys are encrypted and stored in the browser, protected by a passphrase, with a recovery file available. The first validated key claims the instance, unlocking server-side features like agents, skills, and memory, which persist across restarts.

Deployment Options

For users who prefer not to self-host, a hosted version is available at enzo-hub.duckdns.org, which runs the same application code. Additionally, a Google Colab notebook provides a setup option that automates the cloning, dependency installation, UI build, and server boot processes, offering a URL for access from any device, including mobile phones via a Cloudflare tunnel link.