UberLLM

Changelog

2026-10-24 — Webhooks for videos and music

  • Pass a webhook_url when you start a video or music job and we POST the finished job to it once — the moment it ends, succeeded or failed — so you no longer have to poll. Set an optional webhook_secret and every delivery is signed HMAC-SHA256 over the body (X-UberLLM-Signature). Delivery retries with backoff until your URL returns 2xx. See Make a video or image.

2026-10-21 — Prices now include a 10% UberLLM fee

  • Every request (text, images, videos and music) now carries a 10% UberLLM fee on top of the seller's price. Every price on the site already includes it: the price you see is the price you pay. % off list is unchanged. See Pricing.
  • The 5% fee on credit purchases is unchanged. Sellers still pay no fee on their sales.

2026-10-20 — A live market on the home page

  • One models table, on the home page. Prices, % off, liquidity, 24h activity and offers update every 30 seconds on their own, with the market's totals on top: average savings, money saved in 24h and active offers. Click any column to sort it.
  • Tabs for Text, Image, Audio, Video and Decisions. Old /models and /market links open the table.
  • A savings calculator shows what your monthly bill becomes at today's prices.

2026-10-20 — Images, videos and music

  • Make images (/v1/images/generations), videos (/v1/video/generations) and music (/v1/music/generations) with the same key and balance. See Make a video or image.
  • Billed at the real cost of what was made, no markup. A failed, canceled or expired job costs nothing. Jobs show in Dashboard → Activity.
  • Browse them at uberllm.dev, each with its price per video, picture or track.

2026-10-19 — One marketplace, billed at the seller's price

  • Every request now goes to a live marketplace that picks the cheapest healthy seller on each call. You pay exactly what that seller charged — still no inference markup, still a 5% fee on credit purchases only.
  • Automatic retries: a call that fails before any content arrives is retried for you (up to 4 attempts within 45 seconds). A model nobody is serving right now returns 503 with the model named.
  • :floor / :nitro and the provider object are still accepted, but no longer change price or speed. See Models & routing.
  • GET /v1/models lists only models that can be served right now.
  • New accounts start at $0; there is no sign-up credit.

2026-07-24 — Launch

UberLLM is live: a two-sided marketplace for LLM inference. One API key, every model, routed to the cheapest healthy provider.

  • Gateway: OpenAI-compatible /v1/chat/completions and Anthropic-compatible /v1/messages with SSE streaming, automatic failover, and health-aware routing (providers with recent outages are deprioritized, never silently retried forever).
  • Market: 380 models on the book — live-priced curated routes plus external market asks to beat. Modality tabs, capability filters, per-model 24h activity, liquidity, and offer counts at /market.
  • Selling: list any OpenAI-compatible endpoint at your own price (absolute or a multiplier of the reference), pass a one-shot validation probe, earn on every routed request. Passive health scoring only — we never burn your quota with pings.
  • Billing: prepaid credits with a double-entry ledger, per-key spend caps, transparent per-request usage accounting (provider-reported tokens whenever the upstream supplies them). No inference markup — a 5% fee on credit purchases only.
  • Analytics: public market analytics — daily usage and the listed-discount distribution across the book.
  • Docs: drop-in guides for the OpenAI and Anthropic SDKs, Cursor, Cline, Roo, Aider, Continue, LangChain, and the Vercel AI SDK.