Resources
Learn faster. Spend smarter.
Guides, playbooks and answers for getting the most out of real-time inference.
From first call to production confidence
Practical guides for building with AI Inference: prompt patterns, structured output, capability routing, batch jobs, retries and keeping an eye on spend.
Explore →Spend less, ship the same
Practical ways to lower your inference bill: route routine work to lower-priced models, cap max_tokens, stabilise prompts for cache hits and meter retries once.
Explore →Make every response feel instant
How to make inference feel fast: choose responsive models, keep prompts and outputs tight, design for cache hits and measure latency_ms on every response.
Explore →The words that matter, in plain English
Plain-English definitions of the terms you will meet building with AI Inference: tokens, capabilities, routing, fail-over, cache hits, idempotency and micros.
Explore →Answers, right when you need them
Answers to common questions about keys, requests, errors, rate limits, the free allowance and invoices. Still stuck? Email our team and a person will help.
Explore →