FAQ
Minutes: sign up ($1 free, no card), create a key in Console → API Keys, point any OpenAI client at https://api.runaii.cloud/v1. See /docs and /cookbook.
Every account starts with $1 in credits. Modal offers $30/mo compute and Koyeb offers startup credits — ours is smaller but instant, no sales call.
Any OpenAI-compatible client (Python, TypeScript, cURL) plus the Anthropic SDK via /v1/messages. Plus our zero-dep runaii-cli.mjs — see /cli.
Billed per 1M tokens from the model card rate. Cached prompt prefixes bill at the cached rate automatically (up to 97% off). Batch API is 50% off.
API calls return 402 with a top-up link. Set a hard spend limit in Settings and low-credit alerts via email + webhook.
Stripe Checkout (cards) when configured; invoices and custom DPAs on Enterprise. Every charge lands in your ledger — see /console/billing.
Serverless scales to zero and bursts on demand; P50 cold start <800ms on B200 NVFP4. Dedicated deployments stay warm with autoscaling.
60 req/min beta default per workspace, returned as X-RateLimit-* headers with Retry-After on 429. Priority tier raises limits.
Pin GLOBAL / US / EU / APAC per workspace in Settings. EU/APAC carry a 15% premium. VPC/self-hosted on Enterprise.
Yes — dedicated deployments run your checkpoints with per-second billing, or train on runaii (SFT/DPO/RL) and deploy in seconds.