A ANDROMEDA
LOCAL FIRST · OLLAMA · DOCKER MODEL RUNNER · LLAMA.CPP

Your AI assistant.
Your machine.

A capable terminal assistant and a small async agent SDK, in Python and TypeScript. Work with local models, keep your data private, and build at your own pace.

GET STARTED Python: scripts/andromeda · TypeScript: ./start.sh ›
andromeda · ~/my-project LOCAL MODEL
>_ Andromeda (local assistant)
model: gpt-oss:20b /model to change
directory: ~/my-project
memory: local /status for details
› Explain this codebase and trace how a request moves through it.
◦ Reading project files read_file · src/app.py
└ Read 184 lines
• The app has three main parts: request handling, service logic and persistence. The request enters at src/app.py, then moves through the service layer before it reaches the database.
› Ask anything, or type / for commands
auto · shift+tab to cycle · esc to interrupt82% context left
Private by default Your prompts stay on your machine.
Terminal native Built for your workflow
SDK included Build your own agents
Runs locally Ollama, Docker Model Runner, or llama.cpp
CHOOSE YOUR FRAMEWORK

The same agent, in Python or TypeScript.

They share the agent SDK foundations and settings. Pick the one your code lives in; each edition also has its own integrations and setup path.

THE ANDROMEDA HANDBOOK

Everything you need to get moving.

From your first run to building your own agent, find your way around Andromeda.

WHAT'S INSIDE

Everything runs on your machine.

A full assistant out of the box, built from parts you can use on their own.