Infersia
Sign inGet an API key
ModelsPricingDocsStatusTransparency

Documentation

Infersia implements the OpenAI API. If your code already talks to OpenAI, changing the base URL and the key is the entire integration.

base URLtext
https://api.infersia.com/v1

Start here

Quickstart

Make your first request in Python, TypeScript or curl.

Authentication

API keys, spend caps and IP restrictions.

API reference

Every endpoint, parameter and response shape.

Integrations

Cline, Continue, Aider, LangChain, LiteLLM and OpenWebUI.

What’s supported

POST/v1/chat/completions
POST/v1/completions
POST/v1/embeddings
GET/v1/models
GET/v1/key
GET/v1/credits

Request and response shapes match the OpenAI specification exactly, including streaming, tool calling, JSON mode and the usage object. The official openai Python and Node SDKs work without modification — we test against them rather than against our own client.

Where we differ, deliberately

  • Extra response headers. Every response carries x-infersia-quantization, x-infersia-provider and x-infersia-cost-usd, so you can see exactly what served your request and what it cost.
  • Model routing suffixes. model:free selects a free tier, model@provider pins a provider, and model:nitro prefers throughput over price.
  • 402 rather than 429 when out of credit. A prepaid account with no balance gets a clear 402 insufficient_credits with no retry-after, because retrying will not help until you top up.
  • Cost in the usage object. Non-streaming responses include usage.cost in USD.

Getting help

Every response includes an x-request-id. Quote it when you get in touch and we can trace the request end to end — through our gateway to the GPU that served it — without needing anything else from you.

QuickstartGet an API key