Start without commitment, scale with your team, and keep full control of models and token costs.
To build, test and ship on your own terms.
No commitment · Per-minute billing
Idle time never billed
EUR 15 credits · No card required
Usable for LLM tokens and agent runtime · 3 concurrent sessions
Runtime
EUR 0.30/hour
Shared capacity with monthly or annual billing — no per-user pricing.
Monthly billing · No commitment*
1,200 runtime hours included each month
Beyond included hours
EUR 0.25/hour
Everything in Pay as you go, plus:
For high-volume and business-critical deployments.
Annual commitment
Tailored terms, scoped with our team
Custom volume-based runtime pricing
Everything in Team, plus:
*Team remains available with monthly or annual billing. The two-month offer applies only to the first annual term. Monthly billing remains available at EUR 299/month; annual renewals are charged at the then-current annual rate.
Production infrastructure without the systems you would otherwise maintain.
Trusted by teams running production AI agents on Cominty
The same core building tools on every plan. Session, capacity and collaboration limits tailored to each one.
| Pay as you go | Team | Custom | |
|---|---|---|---|
| Capacity | |||
| Runtime | EUR 0.30 / h | 1,200 h / month | Custom |
| Overage rate | EUR 0.30 / h | EUR 0.25 / h | Custom |
| Users | 1 | Unlimited | Custom |
| Concurrent sessions | 20 with Reserve | 200 | Custom |
| Billing | |||
| Per-minute billing · idle never billed | Included | Included | Included |
| Pooled hours across the team | Not included | Included | Included |
| Flexible Reserve (top up) | Included | Included | Included |
| Auto top-up option | Included | Included | Included |
Choose a plan, adjust your usage, and instantly see your runtime, overage and tokens at exact provider price.
Tokens are never marked up. They are drawn from the Reserve at exact provider price, with 0% Cominty fee.
Indicative estimate, excl. VAT.
No subscription. Runtime billed per minute: 350 h at EUR 0.30 / h, plus tokens at exact provider price.
We put transparency at the heart of our solution. Contact us with any questions.
The time your agent is actually running: reasoning, calling tools and executing code in the sandbox. Idle time — waiting for your next message or human approval — is never billed. Usage is metered per minute and rounded up (every started minute is charged), and auto-stop shuts down an inactive agent: you never pay for an agent that sleeps.
EUR 15 of free credits and 3 concurrent sessions, no card required. Install the SDK and your first run ships in four lines of Python.
