Can I sign up and pay right now?
You can sign up now. Sign in with GitHub, open an org, and the Free tier is yours with no card and no expiry. Self-serve checkout for Builder, Pro and Team is built and tested, and card processing is waiting to be switched on, so paid plans start with a quick note for now. Reading the public market record needs no account at all, and never will.
How do one org key and the hosted API quotas work?
Your organization holds one platform key (vitna_key_...) that works across Web APIs, Search and Unblocker. Each service verifies your key against vitna.ai and enforces your plan's daily requests, pages, concurrency and feature access directly in code. Requests past a daily or per-minute allowance return 429 and are never billed as surprise overages.
Is the local model software really free?
Yes. Relay, the Runtime and Anchor are Apache-2.0 and run on your own machines. Your model providers bill your own keys, and the servers or hardware they run on are yours. Serving a request never calls vitna.ai; pushing the runtime's ledger is a separate command you choose to run.
Can I use Vitna's market data in my own product?
Reading it is free, on the pages and through the public API at https://vitna.ai/api/v1, with no key and no quota. Bulk history, archived source documents, export and the right to redistribute are licensed separately, priced per agreement.
What counts as one active workload?
One separately decided unit of work, with its own incumbent, candidates, evidence, constraints and acceptance threshold. An extraction lane and a research-synthesis lane are two workloads even when the same application calls both. Five candidate models inside one lane are still one workload. Pausing or archiving a workload frees its active slot immediately while keeping its full history on record.
What happens if we cancel?
The subscription ends and the org returns to the Free allowance: it keeps every workload it entered, cannot enter another past the first, and hosted API quotas return to Free limits. Nothing else stops. A runtime you are running keeps running, and any signed policy already accepted stays valid on your disk forever.
Do you hold our provider keys, or see our prompts?
No to both. You keep every provider relationship and pay providers directly on your own account. The optional runtime runs on your infrastructure, reads token counts and request shape rather than content, and calls vitna.ai at no point.
Does the workload audit fee credit toward a subscription?
Half of the $10,000 fixed audit fee credits against a new annual Team or Enterprise agreement signed within thirty days of the audit finishing, which is more than a year of Team billed annually at $2,990. The other half is the work itself: the decision brief is yours whichever way the decision goes, and the answer is often stay.
Why are plans never priced per token or per seat?
Because a fee that moves with LLM token flow gives every routing recommendation an incentive problem, and per-seat taxes punish teams for reviewing decisions together. Neutrality that is structural is worth more than neutrality we assert.