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

# API compatibility

> Supported protocols and features.

All models in the current catalog support these endpoints:

| Endpoint | Features |
| - | - |
| `/v1/chat/completions` | Text, streaming, function tools, structured output, reasoning, web search |
| `/v1/responses` | Stateless responses, streaming, function tools, reasoning, web search |
| `/v1/messages` | Messages, streaming, function tools, reasoning, web search and fetch |
| `/v1/messages/count_tokens` | Non-billable input token counting |

Image inputs require an image-capable model. Consult [models](https://docs.otamd.com/models.md) and [images](https://docs.otamd.com/guides/images.md).

Send conversation history explicitly. Stored responses, `previous_response_id`, background jobs and server-managed conversations are unavailable. Client-side tools execute in your application; hosted search and fetch execute as part of inference.

Responses supports function tools. Custom tool formats are not interchangeable with function tools. Check your coding client's supported protocol before connecting.

Published workspace and request policies apply across all models. Requests fail closed when required policies cannot be evaluated.

Use your Limitless API key and `https://trylimitless.ai/v1`. Model availability, active pricing, account credits and rate limits apply to every endpoint.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.