Transparent pricing

Pay per use. Nothing hidden.

Prepaid credit. No subscriptions, no seats, no minimums. Every response tells you exactly what it cost and what's left.

Chat completions

$0.40 /1M tok

OpenAI-compatible completions on MLX.

POST /v1/chat/completions

Embeddings

$0.02 /1M tok

384-dim semantic vectors for RAG, search and clustering.

POST /v1/embeddings

Hosted search · RAG

$0.001 /query

Private per-key vector store + retrieval. No infra to run.

POST /v1/search

Market briefings

$0.05 /call

Live cross-asset signals + an AI analyst narrative.

POST /v1/briefing

What $5 buys you

Concrete, at the rates above — so you can size a budget before writing any code.

Reranking and document upserts are billed at the embedding rate, on the tokens they actually embed — there is no separate per-call fee.

Free credit, no card

Every new key starts with $0.05 of credit, issued instantly by POST /v1/trial. That's ~125,000 chat tokens — plenty to evaluate the API properly. One trial key per network per 24 hours.

Billing FAQ

Are there subscriptions or seats?

No. Prepaid credit only — top up what you want, spend it at the metered rates, nothing recurs.

Does unused credit expire?

No. It stays on your key until you spend it.

How do I top up?

POST /v1/checkout creates a card checkout session; credit lands on the same key. Keys are never rotated on top-up.

What happens when credit runs out?

Calls return 402 insufficient_quota and stop billing — no overage, no surprise invoice. Top up and the same key resumes.

Can I see live pricing programmatically?

Yes — GET /pricing is public, and GET /v1/usage returns your balance and spend.

Start with $0.05 free.

No card. No form. A working key in one request.