Bring your subscription.
Add your Claude, Codex, Gemini, or Kimi credentials. Start with the tools you already use.
YOUR AI, CONNECTED
Turn your Claude, Codex, Gemini, and Kimi subscriptions into one OpenAI-compatible API. Free with no daily cap and no card. Pro is $12 a month.
Encrypted credentials. A key that's yours.
Add your Claude, Codex, Gemini, or Kimi credentials. Start with the tools you already use.
Your credentials are encrypted. You receive the key that unlocks them when you make a request.
Call one API from your apps, scripts, and workflows. Your subscriptions do the rest.
Your side project. Your internal tools. Your next big thing. Put the AI you already use behind the code you want to write.
Pick a provider to explore the API shape.
curl "$PLAN2API_BASE_URL/chat/completions" \ -H "Authorization: Bearer $PLAN2API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "claude/<model>", "messages": [{ "role": "user", "content": "Build something good." }] }'Your subscription credentials are encrypted at rest. Your API key enables decryption when a request arrives, rather than storing the key alongside your credentials.
The design limits exposure from a database-only leak: encrypted credentials alone are not enough to unlock your subscriptions. A stolen API key or a compromised live server is a different risk.
CLIProxyAPI is open source and free, and you host it yourself: you run the server, keep it updated and look after your logins. Plan2API is hosted. You connect your subscriptions in the browser, your logins stay encrypted and only your API key can unlock them, and you get one OpenAI-compatible API with usage, logs and email alerts. Free costs nothing; Pro and Scale lift the limits.
No. Free has no daily cap. It connects 1 subscription and allows 1 request a second, with bursts of 20 and 4 at once, and your subscription's own usage limits are the only cap. When it runs dry, requests get a 429 error that says when it frees up. Pro, at $12 a month, connects 3 subscriptions and moves to the next one from the same provider when one runs dry.
Yes, on Pro and Scale. Connect a Codex subscription, then call /v1/images/generations or /v1/images/edits with any OpenAI SDK. Images render with gpt-image-2 on your own ChatGPT plan and count toward its image limits, which are separate from chat.
Yes. Subscriptions count per connected account, not per provider, so 3 Claude accounts are 3 subscriptions, the same as 1 Claude, 1 Codex and 1 Gemini account. Free connects 1, Pro 3 and Scale 25.
Anytime, from Manage billing in your dashboard. Your plan keeps running until the end of the period you paid for, then your account moves to Free. Stripe handles the payments, invoices and tax as merchant of record.
Your prompts pass through our servers to your provider and back. We never store your prompts, outputs, files or images. For each request we record the time, the model, which key and subscription handled it, token counts, timing and status. If a request fails, we keep the error message, which can quote your provider's response.
You connect a subscription at your own risk, as our terms say. Your provider's terms still apply and may not allow using it through a service like Plan2API. A provider can limit, suspend or close your account, or block access, so a model or a whole provider may stop working here without notice. We may pause a subscription when we see signs it's at risk, and we email you when one gets paused or needs reconnecting.