Docs
Promptoll is OpenAI-compatible. If you know the OpenAI SDK, you know Promptoll.
Quickstart
1. Create an account
Sign up and create an API key from Console → API Keys.
2. Point the OpenAI SDK at Promptoll
from openai import OpenAI client = OpenAI( base_url="https://promptoll.com/v1", api_key="zai_•••", )3. Pick a model — or let us pick
client.chat.completions.create( model="auto", messages=[{"role": "user", "content": "hi"}], )
Routing modes
| auto | Cheapest model in your catalog |
| quality | Highest quality tier available |
| <model-id> | Use this model. If arbitrage is enabled, we may swap for a cheaper equivalent. |
Response headers
| X-Model-Used | Model that actually produced the response |
| X-Latency-Ms | Upstream + routing time in ms |
| X-Cost-Microcents | What we pay upstream (µ¢) |
| X-Price-Microcents | What you are charged (µ¢) |
Streaming
Pass stream=True for SSE. Promptoll forwards every chunk unchanged and bills on the final usage event.