Pricing

reAPI reads the model name in your request, looks at vendor health, current latency, and current load, and hands the call to whichever candidate is most likely to return cleanly fastest. Health checks run continuously, so the route table is always live, not refreshed on a schedule. When the chosen vendor regresses mid-flight, the next call moves automatically and your application never sees the failure.

No. Request bodies and model outputs travel through the AI API aggregator but are never persisted. We retain only billing and audit metadata: model name, token counts, latency, status code, and key identifier. That is what compliance reviews want and what we are willing to be on the hook for. The same posture is the industry baseline among serious gateways.

Set the base URL to https://reapi.ai/api/v1 and supply a key minted in the dashboard. The OpenAI, Anthropic, and Google SDKs continue to work unmodified — reAPI preserves their request and response shapes. Provider-specific paths like Veo or Suno mirror the original vendor request shape, so an existing integration moves over without a rewrite.

Every popular model is fronted by multiple vendors. When one degrades, reAPI moves traffic to the next within a sub-second window and the call still returns. Single-source models surface a structured error rather than guess. The status page publishes a rolling history.

One reAPI key covers every supported model. We handle the vendor relationships behind the scenes, so there are fewer keys to rotate, fewer dashboards to monitor, and a single audit trail across image, video, chat, music, and code workloads. Adding a new provider to the catalog reaches your account without any procurement work on your side.

Pay-as-you-go credits, priced per completed generation — per token, image, or video second depending on the model. There is no subscription and no monthly minimum, credits never expire, and failed generations are refunded automatically. Every model page carries a live pricing table, so you can see the exact rate before you call.

One key reaches 100+ models across image, video, chat, music, and code — Seedance, Seedream, GPT Image, Nano Banana, Veo, Kling, Midjourney, Claude, GPT-5.6, Kimi, Suno, and more. New models reach the catalog within days of vendor release, land behind the same endpoint shape, and show up in your dashboard without any extra setup.

The dispatch decision is made against an always-live route table — vendor health, latency, and load are tracked continuously, so picking a route costs a negligible slice next to model inference time. For generation workloads the model run dominates end to end, and because routing steers around degraded vendors, slow-provider tail latency often drops rather than grows.

Yes. Every new account starts with free credits — no credit card required — enough to make real calls against the cheaper image and video models and to see the dashboard, logs, and per-call costs with your own requests. When you top up later, the pay-as-you-go rate is the same one shown on the model pages.

Direct integrations multiply everything: keys to rotate, dashboards to watch, billing to reconcile, and outages to handle yourself. reAPI collapses that into one key, one bill, and one audit trail, adds automatic failover across vendors, and lets you switch models by changing the model name in the request. The zero-logging posture stays identical across every modality.