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

# Model Families

> The kinds of models you can run with Nito commands, what each is for, and which capabilities are available today.

Models in the Nito catalog fall into a few families grouped by what they do. A family is just a category of model with a shared job: generating text, or reading an image. This page describes the families so you can match a task to the right kind of model.

A term used throughout: a **modality** is a kind of data, such as text or images. A model's input modalities are what it accepts; its output modalities are what it produces. The families below are best understood as combinations of those.

Use the anchors to jump to a family: [Text](#text), [Reasoning](#reasoning), [Vision (image input)](#vision-image-input).

## Text

Text models are the core chat-and-generation family. You send a prompt, you get a generated reply. This is what most Nito commands use when you ask a question or continue a conversation (see [Ask and Continue](/commands/ask-and-continue)). Text models vary in size, speed, context window (how much text they can consider at once), and cost. Because those vary, do not assume two text models are interchangeable on price or context; browse the catalog with the `models` command and use the [decision guide](/models/choosing-a-model).

## Reasoning

Reasoning is a capability of certain text models rather than a family of its own. They work through a problem step by step before committing to a final answer, and they let you control how much effort to spend doing it. Think of it as the difference between answering off the cuff and working the problem through first.

The catalog marks which models support reasoning, so browse with the `models` command and check the entry before relying on it. When a reasoning model is used, the thinking it does is metered separately from ordinary output, so you can see what it cost. See [Reasoning](/features/reasoning).

## Vision (Image Input)

Vision-capable text models accept images as part of a prompt, alongside text, for tasks like describing a photo or reading a chart. Vision here means image input, not image output: the model reads the picture and replies in text. The catalog marks which models support vision, so browse with the `models` command and check the entry. Accepted image formats are PNG, JPEG, and WebP. See [Image Input](/features/multimodal/image-input).

Nito also parses documents (PDF, DOCX, PPTX, XLSX, XLS, and text) so their text can be fed into a prompt with any model. That is the [file parser](/features/server-side-tools/file-parser) at work, not a vision capability: the document is turned into text first, then sent as text. Use vision for pictures and charts; use the file parser for documents.

<Note>
  **Audio and video are not available in the plugin today.** No transcription, speech, or video in or out. Plan around text and vision input.
</Note>

## Common Issues

* **Assuming every text model can see images or reason.** Image input and reasoning are per-model capabilities; the catalog marks which models support each, so check the entry when you browse.
* **Expecting to generate images.** The plugin reads images (vision), it does not create them from a prompt.

## Related Resources

<CardGroup cols={2}>
  <Card title="Choosing the Right Model" icon="magnifying-glass-waveform" href="/models/choosing-a-model">
    A decision guide from use case to model.
  </Card>

  <Card title="All Models" icon="list-timeline" href="/models/all-models">
    Browse every model in the catalog.
  </Card>
</CardGroup>