Built for your stack
Use an OpenAI-compatible endpoint with the tools you already know.
Your AI models,
for less.
One key to build with the models you love.
Pay $10. Get $20 in usage credit.
$0subscription or setup fees
2×usage credit on every top-up
Buy credits
Pay as you go$⌄
You pay $10Save 50% on credit
Pay in ETH · Network fees apply
Pay from your wallet
Top up with SOL on Solana or ETH on Ethereum and Robinhood. Use credit as you go.
Your prompts stay yours
We store usage metadata, never the text of your prompts or responses.
One connection. More possibilities.
OpenAIAnthropicGooglexAIDeepSeekQwen
THE MODEL COLLECTIONThe right mind
Explore all models ↗The right mind
for every idea.
↗
Loading models…
Try a different model name or provider.
Model availability and pricing are shown after provider verification. Unverified models cannot receive requests.
Platform capacity
PREPARING TO LAUNCHweighted tokens in today's operating allowance
Used today0
Verified active models0
Recent purchases
Anonymous · on-chain↗
The first chapter is yours.
Confirmed purchases will appear here. No names or wallet addresses are published.
SMALL CHANGE. BIG POSSIBILITIES.Two lines.
Two lines.
You're connected.
Point your compatible client to Routers.
Your app keeps moving.
.env
# Your new connection
OPENAI_BASE_URL=https://YOUR_DOMAIN/v1
OPENAI_API_KEY=sk-mint-…OpenAI-compatible · Streaming · Model routing
Simple by design.
What am I buying?+
Prepaid dollar-denominated API usage credit, with a separately tracked base-token allowance. Both limits apply.
How do base tokens work?+
Each model consumes base tokens at its configured supplier ratio. One actual model token can consume less or more than one base token. Your dashboard shows both counts.
Can I pay with my wallet?+
Yes. Wallet checkout supports native SOL on Solana and ETH on Ethereum and Robinhood. Credit is added after the server verifies the transaction and its finality. Purchased credit works across all available models, including premium models.
Do you store prompts or responses?+
No. The gateway stores billing metadata: model, token counts, cost, status and timing. The upstream provider processes request content under its own policies.
Can one key use every model?+
Yes. One Routers key works across every currently enabled model available to that key, including GPT, Claude, Gemini, Grok, DeepSeek, Qwen and other listed families. Change only the model ID in your request.