API
Not a REST API you query for your own dashboard data, not yet.
This page describes how CostMyAI actually connects to your AI usage today, and what comes next. The real integration surface is ingestion, built so visibility into your spend never requires handing a third party your credentials.
How CostMyAI connects to your data today
The Verification Engine is the component that makes CostMyAI work without ever holding your provider credentials. It runs inside your own environment, reads your real AI usage and billing data using your own provider keys, which never leave that environment, and pushes neutral, aggregate records to CostMyAI over its own API.
This is the actual integration surface today: an ingestion path, built specifically so visibility into your AI spend never requires handing a third party your credentials.
What it sends
- Aggregate, provider-neutral usage records
- Billed spend, per model and per host
- Token counts, latency, and status metadata
What it does not do yet
- Prompts or completions
- Provider API keys
- Any customer content
- A way to pull your dashboard data back out programmatically
Every number you see in CostMyAI today lives in the dashboard UI. There is no programmatic customer data export yet.
What comes next
The strongest case for a public query API is consistency with what CostMyAI already argues for: a platform built against vendor lock-in should not itself lock a customer's own spend and savings data behind a UI with no way out. A read API letting customers pull their own certified switches, savings, and spend history into their own BI tools or alerting systems is the clearest, most likely next step.
This is not built yet, and it will get built the moment real demand shows it is worth building, not before. If you have a specific integration need today, tell us what you are trying to connect, and what breaks for you without it. That is the single most useful thing you can send us right now.