> ## Documentation Index
> Fetch the complete documentation index at: https://algorand-54fd52b3.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Models

> The catalog and how to choose a model.

On the current Robinhood deployment, Moltworld exposes enabled **chat** models from its configured providers. Each model is named after the provider and model that serve it.

The live list is at [`GET /v1/models`](https://algorand-54fd52b3.mintlify.site/api-reference/endpoints/list-models) and on [motoworld-wif.vercel.app/models](https://moltworldd.vercel.app/models). Robinhood filters out non-chat modalities before they can be quoted or paid.

## Current endpoint

| Modality | Endpoint | Body |
| - | - | - |
| Chat | `POST /v1/models/{model}/chat/completions` | `messages`, `max_tokens`, `temperature` |

Chat requests and responses follow the OpenAI chat completions shape, so existing parsing code works.

## Catalog fields

Each enabled chat model lists its limits and pricing metadata. A body outside the model's limits gets `400` before any wallet authorization.

## Availability

At startup Moltworld checks every model against its provider's catalog. A model whose provider key is missing, or which the provider no longer lists, is withheld and cannot be paid for. `GET /v1/models?include=unavailable` shows withheld models and the reason.

> **Legacy Algorand catalog:** The isolated legacy deployment also documented image, voice, and historical video surfaces. Those endpoints are not enabled on Robinhood and should not be inferred from this page.


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