Pay for tokens, not seats.

No monthly plans. No seats. Use account credits, call any model, and recharge securely with native USDC.

Get API key See every model

How credits work

Three steps. No subscriptions, no auto-renew, no surprise invoices.

Step 1

Start with credits

Add credits. Use one balance across all models.

Step 2

Call any model

Connect your SDK and choose a model.

Step 3

See exact cost

See tokens, latency, and cost for each request.

Recharge status:Credit checkout is not configured on this deployment.Credits are added only after a signed payment confirmation.

Sample rates

Pass-through provider prices with a thin managed-platform margin.

All models

Model	Provider routes	Status	Input / MTok	Output / MTok
Gemini 3.1 Flash-Lite `gemini-3.1-flash-lite`	Google AI Studio	Available	$0.25	$1.50
Gemini 3.5 Flash `gemini-3.5-flash`	Google AI Studio	Available	$1.50	$9.00
Nano Banana Pro `nano-banana-pro`	Google AI Studio	Available	-	$0.134 per image
Veo 3.1 `veo-3.1`

Image and video models are priced per-image or per-second; full breakdown lives on the models page.

One API. Every model.

21 models across 3 providers.

Explore models

One model catalog

Text, image, video, and audio. Connect with your existing OpenAI SDK.

21models

Your choice of route

Choose a provider by price, speed, and availability.

21available routes

Clear usage and cost

Check rates before you build. Track every request in Activity.

Pay as you gono subscriptions

Need dedicated capacity?

For teams that need SSO, on-prem deployment, custom provider adapters, or reserved upstream capacity — talk to us. Custom contract, not a SKU.

Contact sales

Questions, briefly answered.

Are there monthly plans or seats?+

No. Every account is pay-as-you-go. Add prepaid credits with native USDC when you need them; nothing recurs automatically.

How is each call priced?+

Text is per-token in + per-token out. Image is per-image. Video is per-second. We use the upstream provider's published rates with a small managed-platform margin baked in — every request shows its exact cost in Activity.

What happens when credits run out?+

Requests fail with HTTP 402 when balance is below the required request hold. Add balance to continue routing calls.

Can I bring my own provider key?+

Not on the managed plan. 202AI runs the upstream credentials so a single sk-maas-… key gets you Claude + Gemini + OpenAI + MiniMax. For full upstream control, the self-host build is the path.

OpenAI compatibility?+

Drop-in. The /v1/chat/completions endpoint speaks the OpenAI schema across all providers — streaming, tool calls, image inputs translated automatically.

Refunds on unused credits?+