Local preview
Bring your own Vercel AI Gateway key. Provider usage and any provider fees are separate.
- One-command app launch
- Decision control center and receipt inspection
- 45 editable recipes and nine MCP tools
- Scoped agent connections and daily limits
- HTTP API and optional dialogue adviser
Free early access
Use hosted Code Mode or direct MCP tools, a scoped API key, or the playground. We operate the engine and provide Jev access.
- One Jev allowance across MCP, API and playground
- Hosted Code Mode with a separate daily execution limit
- Up to 1 GiB of decision input per UTC day
- Recipes, usage history and revocable access
- No card, waitlist or automatic paid overage
Free access renews at 00:00 UTC each day. Failed calls and calls without reported usage may still count against free access. Up to 10,000 evaluations and 1 GiB of input per account per UTC day; request size and rate limits also apply. Daily capacity may be limited.
What would a paid plan add?
We are exploring larger allowances, saved personal policies and longer history for people who use several agents. A paid plan would buy more capacity and better tools for comparing workflows. Pricing is not finalized, and there is no checkout yet.
What does "cheaper" mean for you?
A smaller decision call can reduce API work if it replaces enough expensive reasoning or filters context without damaging the answer. It can also add overhead. For a fixed-price chat subscription, fewer API-equivalent tokens do not automatically reduce your bill.
Compare complete tasks, including host tool calls, review and escalations. Studio records model attempts and reported tokens; provider invoices remain the authority for charges. Read the evidence before choosing a pattern.