OpenRouter spend, itemized

Connect your OpenRouter organization to Cloud Intelligence™ and see billed cost per model and per hosting provider, next to the rest of your cloud spend.

OpenRouter makes it easy to route LLM traffic to hundreds of models across dozens of providers through one API. That's exactly what makes the bill hard to reason about. Your prepaid credits drain across OpenAI, Anthropic on Bedrock, DeepInfra, Novita, and whatever your team tried last Tuesday, and the only place to see where they went is OpenRouter's own dashboard, cut off from budgets, showback, and everything else your FinOps practice runs on. Every new model provider used to mean either a new integration or a blind spot.

That gap is closed. Cloud Intelligence™ now imports OpenRouter billing natively [CONFIRM: "(available in Early Access)"].

What you get

Daily billed cost per model and per hosting provider, imported into Cloud Analytics as a first-class provider. One connection covers your whole model long tail: if OpenRouter can route to it, the integration reports on it, including the same model served by different hosts. Token usage comes along as the usage metric, with input, output, and cached-token splits available as extended metrics.

The numbers mean what your finance team thinks they mean. Cost is what OpenRouter actually charges you: credits drawn from your prepaid balance. If you bring your own provider keys, the upstream provider bills you directly and OpenRouter only estimates that spend, so the integration keeps the estimate out of Cost (no double counting against your existing AWS or OpenAI feeds) and surfaces it as a separate metric instead.

Because it lands as a regular billing source, everything downstream just works: reports, budgets, anomaly detection, forecasts, and the GenAI Intelligence dashboard, where OpenRouter shows up next to Anthropic, Fireworks AI, and your other AI spend. Per-API-key and per-user daily splits are collected too, ready for chargeback.

Setup is deliberately boring. You create one read-only management key in OpenRouter under Settings, Management Keys, pick a name for the connection, and you're done. Regular inference API keys won't work, and the connection test tells you that immediately rather than failing quietly later. On first connect, Cloud Intelligence™ backfills up to 180 days of history; after that, data refreshes every six hours.

Get started

1. Create a management key in the OpenRouter dashboard under Settings > Management Keys.
2. Connect your OpenRouter organization from Data ingestion and integrations > Integrations.
3. Open the GenAI Intelligence dashboard or build a report grouped by Model and Provider.