Pricing / during the beta

Free while we're still earning it.

Commanda is in development. Every team in the private beta uses it at no cost, on every plan, until we're confident it's worth paying for. The numbers below are what we expect to charge at launch — published early so nobody gets surprised.

Provisional — prices may change before general availability

Solo01

Solo

For individual developers who'd rather direct three agents than type one function.

Freeduring beta

then ~$0 · the single-seat tier stays free

  • Up to 4 concurrent agent runs
  • Editable plans, diff review, full rewind
  • Local context index across your repos
  • Blueprints, hooks and skills
  • MCP servers and editor bridges
  • Bring your own model keys
  • Community support
Start with Solo
Team02 · most teams

Team

For teams who want shared blueprints, shared memory and one place to see what agents shipped.

$39/ developer / month

provisional · free for every beta team

  • Everything in Solo, plus:
  • Up to 12 concurrent runs per developer
  • Shared blueprints, skills and team memory
  • Headless runs in CI on every pull request
  • Org policies: allowlists, spend caps, approvals
  • Ticket and observability integrations
  • Exportable audit traces
  • Priority support, shared channel
Request Team access
Enterprise03

Enterprise

For organisations where "where does the code go?" needs a documented answer.

Talk to us

annual · scoped to your environment

  • Everything in Team, plus:
  • Self-hosted or private-cloud deployment
  • Your own model gateway or local weights
  • SSO/SAML, SCIM provisioning, RBAC
  • Org-wide policy enforcement and reporting
  • Security review and DPA
  • Onboarding and a named engineer
Contact sales
0%margin on model usage. You bring your own keys and pay providers directly.
No seats wastedBilling follows active developers. Dormant seats aren't charged.
Cancel anytimeMonthly plans, no minimum term, and your blueprints stay in your repo.

FAQ / the awkward questions

Answered plainly.

If something here doesn't cover it, write to hi@commanda.ai and you'll get a real answer.

Is it actually usable yet?

It's in private beta and improving weekly. Core loops — fleet, plans, review, rewind — work today on real repositories; some integrations and the enterprise controls are still landing. We'll tell you exactly what's rough before you start, not after.

Do I pay for model usage on top?

You pay your model provider directly with your own API keys, at their prices. We don't resell tokens or take a cut, and we don't meter your usage. If you run local weights, that cost is your electricity bill.

What happens to beta teams when pricing starts?

You'll get notice well in advance, a founding-team discount, and no automatic conversion. Nobody gets billed by surprise for something they were helping us test.

Does my code go to your servers?

No. Indexing, traces and history stay on your machine. Code reaches your chosen model provider only, under your keys — and with a local model, it doesn't leave the machine at all.

Can we self-host?

Self-hosted and private-cloud deployments are part of the enterprise track. If that's a requirement for you, tell us now — early enterprise partners are shaping how it works.

What if we stop paying?

Blueprints, skills and memory are plain files in your repository. Traces export as JSON. Everything an agent produced is already a git commit. There's nothing of yours to hold hostage.

Do you offer discounts?

Yes — free for students, educators, and maintainers of open-source projects, and reduced pricing for non-profits. Email us with a link to your work.

Join the beta while it's free.

Small waves, personal onboarding, and a direct line to the team building it.