Pricing
Pay only for what you use. Start free and pay as you go. Institutional quality, priced for the individual.
Usage is billed in credits, where 1 credit = $1 USD. Credits never expire, and failed requests are refunded automatically — you pay for results, not attempts.
Your AI does not call one endpoint; it chains a few. These are complete answers, priced end to end.
A live scan, then the thread behind the item that matters.
Events on the names you hold, then the full brief on the one that matters.
Official outlooks, results against those outlooks, and one filer's record.
A retrieval-and-reasoning pipeline: ranked winners and losers with cited evidence.
A 15-minute wire poll, 96 times a day. Scheduling is not billed.
Each tool uses one of three models:
- Fixed — a flat fee per request.
- Variable — billed from the result (per development returned, or per 1,000 snippet tokens, etc.).
- Fixed + variable — a per-request fee plus a result-based charge.
Bucketed rates round partial buckets up — 1 to 10 records is one bucket, 11 to 20 is two.
- Alerts cost exactly the endpoint they trigger. Scheduling, deduplication, thresholding, and delivery to email, Telegram, or a webhook are free.
- MCP and plugins cost exactly the tool your AI invokes. Using Ultralayer from ChatGPT, Claude, or any other AI costs the same as calling the API yourself.
- The app draws on the same balance, at the same rates. One balance covers every surface.
- Failed requests are refunded automatically.
There is no seat fee, no data-feed licence, no annual contract, and no minimum spend.