Plans for individuals, teams, and governments.
Every plan is a key, and one key covers both: the agent in your terminal and the OpenAI-compatible API. Pick the usage that fits - from one person to a sovereign deployment in your jurisdiction.
Pro
Individuals
per person, per month
Access to Cosmic 1 for individual daily use through the API.
- Cosmic 1
- API access
- OpenAI-compatible API key
- 8,192-token context window
- Standard usage limits
No payment is taken yet.
Team
Most popularTeams and organizations
per seat, per month, minimum 3 seats
Everything in Pro, with higher usage and shared administration across your seats.
- Everything in Pro
- 5x Pro's API rate limit
- Shared admin console
- SSO and role-based access
- Usage dashboards
- Priority support
3-seat minimum. No payment is taken yet.
Ultra
Power users
per person, per month
Higher usage limits and priority access for heavy individual workloads.
- Everything in Pro
- 20x Pro's API rate limit
- Priority access at peak times
- Early access to model updates
- Longer autonomous agent runs
- Direct support
No payment is taken yet.
No payment is taken. Choosing a plan records your choice and nothing else. There is no card to enter and nothing is billed. We record the plan you pick so we can shape usage limits around it, and the prices above are what each plan will cost.
Governments & regulated infrastructure
Sovereign
A dedicated, isolated instance in your own jurisdiction, run to your controls - with zero data retention and a contractual data boundary. Priced on term and scope.
In every plan
- Streaming and tool calls
- OpenAI-compatible API
- Keyed, logged, and revocable
What each plan includes.
Every plan includes Cosmic 1. Higher plans add more usage and more administrative controls.
Requests
The figures each plan states and the routes enforce. They are requests, per account - nothing is metered in tokens.
| Limit | Pro | Team | Ultra |
|---|---|---|---|
| Dashboard chatMessages sent from the chat on this site. | 40 | 60 | 120 |
| API, per minuteRequests to /v1 with an API key. Each key also has its own bucket at this figure. | 60 | 300 | 1,200 |
| API, per dayRequests to /v1 across every key, in UTC days. | 2,000 | 9,000 | 40,000 |
| This sessionThe chat and the API together. Everyone shares one model host, and this is your share of a sitting. | 400 | 900 | 2,000 |
| This weekThe chat and the API together, over a rolling seven days. | 6,000 | 12,000 | 30,000 |
Team’s daily figure is per seat: 3,000 requests a day for every seat on the account, shown above at its 3-seat minimum. Going past a limit answers 429 with a retry-after header, never a bill.
What each plan includes
| Feature | Pro | Team | Ultra |
|---|---|---|---|
| Agent and API accessone key, Cosmic 1 | Included | Included | Included |
| OpenAI-compatible APIchat, streaming, tools | Included | Included | Included |
| Priority access at peak times | Not included | Included | Included |
| Central billing and adminseats and roles | Not included | Included | Not included |
| SSO and role-based access | Not included | Included | Not included |
| Early access to model updates | Not included | Not included | Included |
| Support | Community | Priority | Direct |
Every plan includes Cosmic 1. A Sovereign deployment adds isolation in your jurisdiction, a contractual data boundary, and zero data retention.
Common questions about plans.
How do the usage limits work?
They are request limits, not a monthly token allowance - nothing here is metered in tokens. Each plan states three figures, all per account: 40 requests a minute in the dashboard chat, and on the API 60 requests a minute and 2,000 requests a day. Those are Pro's; Team is five times Pro on the API and Ultra twenty, and Team's daily figure grows with its seats. A request past a limit answers 429 with a retry-after header rather than a bill.
Can I use the API directly?
Yes. Every plan comes with an OpenAI-compatible API key. Point any OpenAI SDK at the Ultrac base URL and use the cosmic-1 model ID.
How does Team pricing work?
Team is priced per seat, per month, with central administration and the daily API allowance pooled across the whole team - 3,000 requests a day for every seat on the account. There is a 3-seat minimum. Choosing Team records the plan and the seat count you asked for. No payment is taken yet.
Do you offer a government or air-gapped option?
Yes. For a dedicated, isolated instance in your own jurisdiction - in-region or government cloud - with zero data retention, a contractual data boundary, and audit logging, talk to us about a Sovereign deployment. It reaches the same model over the same API, priced on term and scope.
Something else? access@ultrac.io
Start with API access. Add a Sovereign deployment when required.
Ultrac is invite only. Join the waitlist for access, or request a Sovereign briefing for governments and regulated infrastructure.
access@ultrac.io · SpinalFluid