> ## Documentation Index
> Fetch the complete documentation index at: https://docs.idem0.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# idem0

> Exactly-once for your LLM API calls. Retry safely, replay any call, never double-pay.

## What idem0 is

idem0 is a proxy that sits in front of the Anthropic and OpenAI APIs and makes your calls **idempotent**: give a call a stable key, and running it again returns the **exact same response** you got the first time — instantly, for **zero credits**.

You keep using the official provider SDK. idem0 is three lines of config.

## The problem it kills

Your agent calls a model. The request times out, the process crashes, a queue redelivers the job, or a retry fires — so the call runs **again**. You pay twice, and you might get a **different** answer the second time. Across an agent loop that's real money and real nondeterminism.

With idem0, the second call is a **replay**: same bytes, same result, no model run, no charge.

## How it works

* Attach a stable `Idempotency-Key` to a request — a job id, a run step, anything you own.
* The first call runs for real and idem0 stores the result.
* Any later call with the same key returns the stored result, tagged `idem0-idempotent-replayed: true`, for **0 credits**.

<Note>
  **BYOK.** Your provider key stays in your provider SDK, on its normal `Authorization` / `x-api-key` header. idem0 never sees it, stores it, or logs it.
</Note>

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    First call, then watch a replay come back for free — under 5 minutes.
  </Card>

  <Card title="Concepts" icon="book" href="/concepts">
    Idempotency keys and replay, briefly.
  </Card>
</CardGroup>