Pricing
The Prices page lets you subscribe to a plan and top up tokens through Stripe Checkout. Clicking a plan's call-to-action opens Stripe's hosted checkout; on return you see a success or cancellation toast.
Subscription plans
Recurring plans, each shown as a card with a feature checklist and a subscribe button. The current plan is marked Current; the top tier is "Contact sales" rather than a self-serve checkout. Typical tiers:
| Plan | Positioning |
|---|---|
| Free (trial) | Starter quotas to try the platform |
| Discoverer | Entry paid tier |
| Standard | Mid tier — larger quotas |
| Premium | High tier |
| Enterprise | Contact sales — SSO, private instances, security features |
Each card lists the plan's quotas — Disk Space, Event Subscriptions, Webhooks count, API Endpoint Count, API Queries/Month, and the per-query cost of additional API calls above the quota. Cards collapse the longer feature list behind a Show N more toggle.
The numbers shown on the cards are the live source of truth — this page reflects your account's current plan and the offers available to it.
Token packages
Below the subscriptions, one-time token packages can be bought on top of any subscription — a top-up that draws from the same balance your API usage consumes. These use a one-time Stripe payment (button label Buy) rather than a recurring subscription.
How usage is billed
Everything bills from a single token pool — plain API requests and platform automation alike:
- API requests consume tokens per call, within your plan's monthly quota; calls above the quota cost the per-query rate shown on your plan card.
- AI Agents, Skills, and VM runtime are billed from the same quota, each action type carrying its own multiplier (an agent step, a VM start, VM runtime minutes, and non-VM skill calls each cost a defined amount of tokens).
- Bring your own AI key: the agents' own LLM calls — made with your own AI Credentials — are not marked up or metered by the platform. Only the orchestration (agent steps, skill calls, VM runtime) is billed.
Plans also carry hard operational limits that are enforced per subscription: Max Agent Depth, Max Agent Execution Time, Max Parallel VMs, and VM Runtime per month. Hitting one stops the run with a plain-language reason rather than an unbounded bill.
Related: