Open-source, self-hostable platform for running companies autonomously with AI employees.

About

Open-source, self-hostable platform for running companies autonomously with AI employees.

A Company has human Members and AI Employees. Every AI Employee has a Soul (their constitution), a set of Skills (markdown playbooks), and Routines (scheduled, cron-driven work) — all stored in the database. Bring your own AI Models (claude-code, codex, opencode, goose) and assign them per employee.

Disclaimer: Some parts of this software are AI generated. Use at your own risk. It is open source and provided without warranty of any kind — see LICENSE.

Repo layout

genosyn/
├── App/         # Product app — Express + TypeORM + React + Vite + Tailwind
├── Home/        # Marketing site — React + Vite + Tailwind
├── CLI/         # `genosyn` cluster-maintainer CLI (bash)
├── AGENTS.md    # Instructions for AI coding agents (read this first)
├── ROADMAP.md   # Vocabulary, milestones, backlog
└── CLAUDE.md    # Pointer to AGENTS.md

Quickstart

One-command install (recommended)

curl -fsSL https://genosyn.com/install.sh | bash

This installs the genosyn CLI to /usr/local/bin (falls back to ~/.local/bin without sudo), pulls the latest image, and starts a container on port 8471 with a persistent genosyn-data volume. Re-run it any time to upgrade.

Day-to-day management:

genosyn status           # show running state, image, volume, port
genosyn upgrade          # self-update the CLI, pull latest image, recreate
genosyn self-upgrade     # update only the genosyn CLI script
genosyn logs -f          # tail server logs
genosyn backup           # tarball the data volume
genosyn restore <file>   # restore a backup
genosyn prune            # remove orphaned images from prior upgrades
genosyn stop | start | restart
genosyn uninstall [--purge]
genosyn help             # full command reference

Defaults can be overridden with flags (--port, --name, --volume, --image) or env vars (GENOSYN_PORT, GENOSYN_NAME, GENOSYN_VOLUME, GENOSYN_IMAGE).

Docker (manual)

Pre-built images are published to GitHub Container Registry on every push to main.

# Product app — API + UI on one port
docker run -d --name genosyn -p 8471:8471 \
  -v genosyn-data:/app/data \
  ghcr.io/genosyn/app:latest
# → http://localhost:8471

# Marketing site
docker run -d --name genosyn-home -p 8472:3000 \
  ghcr.io/genosyn/home:latest
# → http://localhost:8472

The genosyn-data volume holds the SQLite database (Soul, Skill, Routine, and Run content live there) plus per-employee provider credentials on disk. It persists across restarts and image upgrades.

To override config.ts without rebuilding, mount your own on top:

docker run -d --name genosyn -p 8471:8471 \
  -v genosyn-data:/app/data \
  -v "$PWD/config.ts:/app/config.ts:ro" \
  ghcr.io/genosyn/app:latest

From source

# Product app — API + UI on one port
cd App && npm install && npm run dev
# → http://localhost:8471

# Marketing site
cd Home && npm install && npm run dev
# → http://localhost:8472

See App/README.md and Home/README.md for production builds, config, and scripts.

Contributing

Read AGENTS.md before opening a PR — it covers the project vocabulary (Routine, Soul, Skill, AI Employee), the stack, and code conventions. Check ROADMAP.md for what's in flight and what's next.

License

MIT © HackerBay, Inc.