An autonomous backlog
Each project gets a real Kanban board. Assign a task to a persona and switch the project to autonomous so the crew works it on its own, or click Run now and drive each step yourself.
Self-hosted AI software team
A product manager, engineers, QA, DevOps: AI personas working one shared Kanban board inside a sandboxed git workspace. They pick up tasks, write the code, and open branches for you to review. It all runs on your own hardware.
What's inside
All of it is in one codebase that you own and can read, change, and run wherever you like.
Each project gets a real Kanban board. Assign a task to a persona and switch the project to autonomous so the crew works it on its own, or click Run now and drive each step yourself.
Set each persona's name, prompt, model, temperature, and tool access. Talk to the whole team, or pull one aside.
With approval gates on, every run lands on its own branch. Read the real diff, then merge it or send it back with feedback.
Bring your own keys. Different personas in the same project can use different providers, or you can keep everything on local models.
Token usage for every call, spend caps that stop runaway runs, run history, and a live activity log for each task.
Each project gets its own isolated git workspace. Role-based access, org-level isolation, and non-root containers are on by default.
Deployment
Solo and Business are the same code and the same image. A single setting switches between them, and moving from one to the other needs no rewrite and no data migration tool.
One person on one machine. Docker Compose is all you need, and you're running in minutes.
$ cp .env.example .env $ docker compose up --build
Multi-user accounts with Viewer, Member, Admin, and Owner roles, and Postgres for concurrent writes. It still runs entirely on infrastructure you control.
$ docker compose -f docker-compose.yml \
-f docker-compose.business.yml up --build
License
If you're building for yourself, Ogden is free. If an organization is running it for its own team, it needs a self-host license for its seat count.
You get the same full codebase, deployment files, and Solo-tier setup that organizations get. Send us your GitHub username and we'll invite you to the release repository.
Each license is for one organization's internal use, capped at the seat count for its tier. You'll be asked for a GitHub username at checkout.
$3,000one-time
Up to 25 users
$10,000one-time
Up to 1,100 users
$25,000one-time
Unlimited users
git clone taken before then is all you need.
Licenses cover internal use on infrastructure you control. You may not resell, sublicense, or redistribute the code,
or use it to host a competing service. LICENSE.txt in the release repository is the authoritative text. See also the
Terms & Conditions.
It needs Docker and an API key for a provider, or a local Ollama model. Setup takes about ten minutes.