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

# Getting started

> Install Fella, connect a model, open a folder, and inspect a tool-backed answer.

# Get an inspectable answer

By the end of this guide, Fella will have scanned a folder, answered a question from its contents, and shown you the computation behind the result.

<Steps>
  <Step title="Install the desktop app">
    Download the current unsigned build from [GitHub Releases](https://github.com/Avijit-Kumar-GIT/fella/releases/latest). Releases provide a universal macOS `.dmg`, Windows installers, and x86\_64 Linux `.AppImage` and `.deb` artifacts.

    ```bash filename="macOS or Linux" theme={null}
    curl -fsSL https://lilfella.app/install.sh | sh
    ```

    ```powershell filename="Windows PowerShell" theme={null}
    irm https://lilfella.app/install.ps1 | iex
    ```

    Current builds are unsigned. Review a script before piping it to a shell, or download an installer and compare it with the release's `SHA256SUMS`.
  </Step>

  <Step title="Connect a tool-capable model">
    Fella is BYOK-only. Connect a hosted provider with your own API key before
    asking a question.

    ```bash filename="terminal" theme={null}
    /login ollama-cloud key [api-key]
    /model gemma4:31b
    ```

    Run `/login`, choose OpenAI, Vercel AI Gateway, xAI, Ollama Cloud, OpenRouter,
    or a custom OpenAI-compatible endpoint, then use `/model` to inspect the
    provider's current model list.

    ```text filename="fella commands" theme={null}
    /login
    /model
    ```

    Hosted providers receive prompts and tool results. Keys are stored in local `auth.json` with `0600` permissions where supported, not in the settings database, browser storage, or transcript.
  </Step>

  <Step title="Open a focused folder">
    Choose **Open folder**, drag a folder into the window, or enter:

    ```text filename="fella command" theme={null}
    /open /path/to/your/files
    ```

    or to open the file explorer:

    ```text filename="fella command" theme={null}
    /open
    ```

    Start with a small folder containing only relevant files. Fella scans up to eight levels deep, skips hidden paths, does not follow links, and applies root `.fellaignore` name or path-prefix rules. A root `fella.md` supplies context and is not treated as evidence data.
  </Step>

  <Step title="Confirm what loaded">
    Run `/files`.

    ```text filename="fella commands" theme={null}
    /files
    ```

    Check skipped files, inferred types, units, date shapes, nulls, and ingest notes before trusting an aggregate. This catches the most common source problems earlier than a natural-language question does.
  </Step>

  <Step title="Ask and inspect">
    Ask for a bounded outcome that names the measure and time range.

    ```text filename="first question" theme={null}
    Using transactions.csv, compare total grocery spending in January and February 2025. Show both totals and the difference.
    ```

    Open the **working** fold under the response. Confirm the right source was selected, the query filters match your wording, units are correct, returned rows support the answer, and no verification warning needs attention.
  </Step>
</Steps>

## What the shipped app can use

| Capability        | Current support                                                                 | Qualification                                                                         |
| ----------------- | ------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------- |
| Desktop platforms | macOS 10.15+, Windows x64, Linux x86\_64 with WebKitGTK 4.1                     | Published macOS build is universal; Windows and Linux release jobs target x64 runners |
| Delimited tables  | `.csv`, `.tsv`, `.tab`                                                          | Loaded into SQLite; up to 2,000,000 rows per delimited file by default                |
| JSON tables       | `.json`, `.ndjson`, `.jsonl`                                                    | Expected to contain tabular records                                                   |
| Spreadsheets      | `.xlsx`, `.xlsm`, `.xlsb`, `.xls`                                               | Each usable sheet becomes a table                                                     |
| Documents         | `.pdf`, `.txt`, `.text`, `.md`, `.markdown`, `.log`                             | PDFs need an extractable text layer; there is no OCR                                  |
| Parquet           | `.parquet`, `.pq`                                                               | Requires a custom `--features duckdb` build; not in release builds or CI              |
| Built-in tools    | File listing, table inspection, read-only SQL, text search/read, Python, charts | The model chooses tools; general answers can use none                                 |
| Charts            | Bar and line                                                                    | Maximum 12 categories and 2 series                                                    |

SQL result materialization is capped at 1,000 rows by default and a query is interrupted after 15 seconds. These are guardrails, not target dataset sizes.

## A good first-question pattern

```text theme={null}
Using [source], calculate [measure] for [scope].
Compare/group it by [dimension]. Show [specific outputs].
```

However, the goal of fella is to eventually support looser question patterns such as

```
Did my spending go up or down from 2023 to 2024? How much and what caused it?
```

Specific wording reduces interpretation errors. If the answer matters, inspect the query rather than relying only on a green verification label.

## Next steps

<CardGroup cols={2}>
  <Card title="Use task recipes" icon="list-checks" href="/examples">
    Adapt tested prompt shapes to your own files.
  </Card>

  <Card title="Read the limits" icon="triangle-alert" href="/limitations">
    Understand verification, Python, formats, and scale before relying on a result.
  </Card>
</CardGroup>