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

# Interface Overview

> How the TokenDog API reference is organized, and integration guidance

## Usage overview

This page explains how the TokenDog API reference is organized and how to approach integration. The reference has three parts — **Guides**, **Endpoints**, and **Examples**. Read this page first, then jump as needed.

## Basic integration info

* The unified `BaseUrl` is `https://tokendog.io`.
* To call an endpoint, append its path to the base URL, e.g. `https://tokendog.io/v1/chat/completions`.

## Navigation overview

<CardGroup cols={3}>
  <Card title="Pick an API format" href="/en/api-reference/examples/basic-chat">
    **Start Here** — decide which API format you'll use, then go to the matching endpoint page instead of hopping between unrelated APIs.

    You'll find: OpenAI-compatible → `/v1/chat/completions`, `/v1/responses`; Claude → `/v1/messages`; Gemini → `/v1beta/...:generateContent`.
  </Card>

  <Card title="Understand rate limits" href="/en/api-reference/guide/rate-limits">
    **Limits** — treat request rate, concurrency, and output size as engineering constraints, not an afterthought.

    You'll find: request windows, concurrency caps, over-limit handling.
  </Card>

  <Card title="Locate the failure" href="/en/api-reference/guide/error-codes">
    **Debug** — the error-codes page helps you tell quickly whether a failure is auth, params, rate limit, or server-side.

    You'll find: auth failures, parameter validation, server-side errors.
  </Card>
</CardGroup>

## Suggested reading order

<Steps>
  <Step title="Pick the API format">Decide which format you'll use: OpenAI-compatible, Claude, or Gemini.</Step>
  <Step title="Integrate the endpoint">Go to the matching endpoint page and build straight from the request structure.</Step>
  <Step title="Add engineering constraints">Finish with the rate-limit and error-code docs for constraints and debugging.</Step>
</Steps>

## Integration tips

* Confirm requests go to the unified `BaseUrl` `https://tokendog.io`, then append the specific endpoint for your chosen API format.
* If you already know you're using OpenAI, Claude, Gemini, Zhipu GLM, or MiniMax, go straight to the matching endpoint pages.
* If you're still mapping out your flow, read the examples first, then return to the endpoint pages to implement.
* If you're building a unified platform layer, converge provider differences in a server-side gateway.