OpenAI compatible. Base URL:
POST /v1/chat/completions or POST /api/v1/chat/completions
List the models you can use with GET /v1/models. Some models use more tokens per reply, shown as cost. Browse them at /models.
curl /v1/chat/completions \
-H "Authorization: Bearer flame_sk_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"MODEL_ID","messages":[{"role":"user","content":"Hello"}]}'
import requests
r = requests.post("/v1/chat/completions",
headers={"Authorization": "Bearer flame_sk_YOUR_KEY"},
json={"model": "MODEL_ID", "messages": [{"role": "user", "content": "Hello"}]})
print(r.json()["choices"][0]["message"]["content"])
Create an account at /plans, then press Create key. The key is shown once, so copy it. All your keys share your account balance.
Every reply uses tokens from your key's balance. When the balance reaches zero the API answers 402 insufficient_quota. Check it any time:
GET /api/v1/account returns balance, unlimited, total_tokens, orders
GET /api/v1/packs packs on sale (no key needed)
POST /api/v1/orders {"pack_id":"builder"} returns UPI payment details
POST /api/v1/orders/{id}/utr {"utr":"123456789012"}
Each request is capped at a maximum number of output tokens set by the owner, and keys have a per-minute limit. Streaming is supported with "stream": true.