GlianaAI — pay-per-call AI inference across 60+ models

# GlianaAI > Pay-per-call AI inference across 90+ generative models (image, video, audio, music, speech) behind one HTTP endpoint. Each request returns HTTP 402 with a price and your client pays from its own wallet (USDC over MPP/Tempo, x402 on Base, Solana, or BNB Smart Chain), then retries — no signup and no API key. Callers who cannot settle from a wallet can instead hold a prepaid balance and send an API key; same endpoints, same prices. Part of Gliana Labs. ## How it works - Call `POST https://api.glianalabs.com/v1/infer` with `{ model, prompt, ... }`. - Unpaid → `402 Payment Required` with the price and accepted rails. - Client pays from its own account and retries with the payment credential → model runs → output returned. - No account, API key, or subscription. Payment is the authentication. - Alternative: sign in at https://dashboard.glianalabs.com, top up a prepaid balance, and send `Authorization: Bearer gk_...` instead of paying per call. ## Payment rails - MPP (Tempo + Stripe) — the primary rail (`mppx` client). - x402 — USDC on Base and Solana (`/x402/<model>`). - b402 — gasless USDT/USDC/USD1 on BNB Smart Chain (`/b402/<model>`). - Pay-as-you-go session channels for many calls (`/v1/session/infer`). ## Models 90+ models across image, video, image-to-video, video-to-video, image-to-image, text-to-speech, speech-to-text, music, and LLM chat. Also flat-priced utility APIs (scraping, screenshots, social card rendering, video summarisation, face matching, blockchain data) and multi-model recipes billed as one call. Live catalog and prices: - Models: https://ai.glianalabs.com/models - Pricing (single source): https://api.glianalabs.com/v1/price?model=<id> - Full catalog with per-call prices: https://api.glianalabs.com/v1/models - Every endpoint, machine-readable: https://api.glianalabs.com/openapi.json - Catalog API: https://api.glianalabs.com/v1/models ## Capability + tool endpoints - OpenAI-style (model in the body): `POST /v1/im