One prepaid key for your AI coding tools
Kredo gives you API access to leading open models at plain per-token prices. Add credit, create a key, and use it from Codex, Claude Code, Cursor or any OpenAI-compatible SDK. You pay only for the tokens you use.
- Step 1
Add credit
Top up your balance. There is no subscription and no monthly fee; credit is only used when you make requests.
- Step 2
Create an API key
Make a key for each tool or machine. You can revoke any of them at any time.
- Step 3
Point your tools at it
Set the base URL and key in Codex, Claude Code, Cursor or any OpenAI-compatible SDK. Streaming works as usual.
Pricing
Prices are in US dollars per one million tokens. Cached input is the part of a prompt the model has already seen recently, billed at a lower rate.
| Model | Model ID | Context | Input | Cached input | Output |
|---|---|---|---|---|---|
| DeepSeek V4 Pro | deepseek-v4-pro | 128K | $0.449 | $0.047 | $1.347 |
| GLM-5.3 | glm-5.3 | 128K | $1.428 | $0.334 | $4.66 |
| Qwen3-VL Plustext + images | qwen3-vl-plus | 128K | $0.204 | same as input | $1.652 |
| Qwen3.8 Flash | qwen3.8-flash | 128K | $0.161 | $0.017 | $0.543 |
| Qwen3.8 Maxtext + images | qwen3.8-max | 128K | $2.264 | $0.284 | $6.79 |
| Qwen3.8 Max Rawtext + images | qwen3.8-max-raw | 128K | $2.264 | $0.284 | $6.79 |
How billing works
- Prepaid. Requests draw on your balance. When it runs out, requests are refused until you add credit; you can never go into debt.
- Held, then charged. Before a request starts we set aside the most it could cost. When the response finishes we charge the real cost and free the rest straight away.
- No charge for failures. If the model provider returns an error, the amount set aside goes back to your balance.
A first request
curl
curl https://kredo.firebirds.cc/v1/chat/completions \
-H "Authorization: Bearer $KREDO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "deepseek-v4-pro",
"messages": [{"role": "user", "content": "Say hello in one sentence."}],
"max_tokens": 200
}'