Run any model.
Wire them together.
One metered catalog behind a playground, an async developer API, and a node canvas you can publish as its own endpoint. Same models, same billing, three surfaces.
One contract for every model
Every generation endpoint returns a request_id and a poll url, then fires a webhook. No endpoint blocks, no per-provider retry logic, and the same request record whichever model you called.
- Five timing phases on every request, so you can see whether it was the queue or the model
- Credits reserved on submit and settled on completion, to the cent
- Concurrency that scales with what you spend — no tier to negotiate
bash
curl -X POST https://api.ezflows.io/v1/models/gpt-image-2 \
-H "Authorization: Bearer $EZFLOWS_API_KEY" \
-H "Content-Type: application/json" \
-d '{"prompt": "a wide shot of a desert at dusk"}'
# { "request_id": "req_8f2a41c9", "poll_url": "...", "status_url": "..." }Playground
Pick a model, tweak parameters, and watch the price update as you go. The cost is on the run button before you press it.
Developer API
The same catalog behind one async, metered contract. Copy-paste curl for every endpoint.
Node canvas
Wire models into a pipeline and publish it. One action serves an API endpoint for developers and a form for everyone else.