LLM credits for you and your bots, paid from Bitkit.
One OpenAI-compatible API for GPT, Gemini, DeepSeek, Kimi and MiniMax. Sign in with Pubky, top up in USDT, no card and no KYC. You pay for the tokens you use.
Sign in with Pubky
Scan a code with Pubky Ring. No email, no password: LLMKY only gets your public key.
Top up from Bitkit
5, 10, 20 USDT on Arbitrum. The request lands in your Bitkit, you approve, the credits are there.
A key per bot or app
One key per bot or app, each with its own daily limit. They all spend your one balance, and you can cut any of them off in a click.
Prices
In USDT per million tokens (about 750 000 words). Each call is billed from the tokens the model reports, to the micro-USDT. Credits do not expire.
| Model | Input | Output | API id |
|---|---|---|---|
| Eco | |||
| GPT-5 nano | 0.057 | 0.460 | gpt-5-nano |
| DeepSeek V4 Flash | 0.161 | 0.322 | deepseek-v4-flash |
| GPT-5 mini | 0.287 | 2.30 | gpt-5-mini |
| Gemini 3.5 Flash-Lite | 0.345 | 1.38 | gemini-3.5-flash-lite |
| MiniMax M2.5 | 0.345 | 1.38 | MiniMax-M2.5 |
| Gemini 2.5 Flash | 0.345 | 2.88 | gemini-2.5-flash |
| Standard | |||
| Kimi K2.5 | 0.690 | 3.45 | kimi-k2.5 |
| GPT-5 | 1.44 | 11.50 | gpt-5 |
| DeepSeek V4 Pro | 2.00 | 4.00 | deepseek-v4-pro |
Provider price + 15 %. Cached input is billed at the provider's cached rate where it has one.
For your bots and apps
Any OpenAI client works: point it at LLMKY and use the key of your bot or app.
curl https://<llmky-host>/v1/chat/completions \
-H "Authorization: Bearer llmky_…" \
-H "Content-Type: application/json" \
-d '{"model": "gpt-5-nano",
"messages": [{"role": "user", "content": "Hello"}]}'