> ## 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.

# File Parser

> Attach a document with --file and Nito reads it for you, extracting the text so any model can answer questions about it.

When you attach a document, Nito extracts its text so the model can read it. The document becomes text before the model sees it, so this works with any model you pick.

## Attaching a Document

Add a document to your prompt with the `--file` flag, then ask your question. The flag is repeatable, so you can attach several documents in one prompt.

<Tabs>
  <Tab title="Claude Code">
    ```text theme={"theme":{"light":"github-light","dark":"github-dark"},"languages":{"custom":["typescript","python","curl"]}}
    /nito:ask --file report.pdf Summarize the key risks in this document.
    ```
  </Tab>

  <Tab title="Codex">
    ```text theme={"theme":{"light":"github-light","dark":"github-dark"},"languages":{"custom":["typescript","python","curl"]}}
    $nito:ask --file report.pdf Summarize the key risks in this document.
    ```
  </Tab>
</Tabs>

You can also attach documents from your terminal with `nito ask --file ...`. See [Ask and fusion from the terminal](/cli/ask-and-fusion-from-terminal).

## Supported Document Kinds

Nito reads text and common office documents:

* PDF
* DOCX
* PPTX
* XLSX
* XLS
* Plain text

Images are not documents. A picture goes to a vision-capable model instead of the parser, so send it down the image path described in [Image input](/features/multimodal/image-input).

## Works With Any Model

Because the document is parsed into text before the model runs, every model can read what you attach. You do not have to choose a special document model. Pick whatever model fits the task (see [Choosing a model](/commands/choosing-a-model)), and the extracted text rides along in context.

## Worked Example

Attach a document and ask a pointed question. Here we pull specific fields out of a small report using the default model:

<Tabs>
  <Tab title="Claude Code">
    ```text theme={"theme":{"light":"github-light","dark":"github-dark"},"languages":{"custom":["typescript","python","curl"]}}
    /nito:ask --file report.txt What is the top region and the Q3 risk?
    ```
  </Tab>

  <Tab title="Codex">
    ```text theme={"theme":{"light":"github-light","dark":"github-dark"},"languages":{"custom":["typescript","python","curl"]}}
    $nito:ask --file report.txt What is the top region and the Q3 risk?
    ```
  </Tab>
</Tabs>

Nito extracts the report's text and the model answers from it, returning the top region and the Q3 risk straight out of the document. The same flow works for summarizing a PDF, listing the deadlines in a contract, or reading a spreadsheet.

## Related Resources

<CardGroup cols={3}>
  <Card title="How Nito Works" icon="gears" href="/get-started/how-it-works">
    Understand how Nito handles your prompt.
  </Card>

  <Card title="PDF and Document Input" icon="file-pdf" href="/features/multimodal/pdf-input">
    Attaching PDFs and office documents in practice.
  </Card>

  <Card title="Image Input" icon="image" href="/features/multimodal/image-input">
    Send a picture to a vision-capable model instead.
  </Card>
</CardGroup>