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

# Speech to Text (Faheem)

> Wittify's Arabic-first speech recognition and transcription engine. Turn audio into editable, speaker-aware text in 16 languages.

<Note>
  **Faheem** turns audio into editable text. Upload a recording or speak straight into your browser, get a speaker-aware transcript with word-level timestamps, edit it inline, and export to TXT, DOCX, PDF, SRT, or VTT.
</Note>

## What you get

<CardGroup cols={2}>
  <Card title="Arabic-first" icon="globe">
    Built for Arabic and its dialects. Works with 16 languages total: Arabic, English, French, Spanish, German, Italian, Portuguese, Russian, Turkish, Persian, Hindi, Urdu, Indonesian, Malay, Hebrew, and Chinese.
  </Card>

  <Card title="Speaker diarization" icon="users">
    Automatically tells who said what. Rename, merge, or split speakers in the editor.
  </Card>

  <Card title="Word-level timestamps" icon="clock">
    Every word carries a timestamp. Click any word in the transcript to jump the audio player to that exact moment.
  </Card>

  <Card title="Live transcription" icon="microphone">
    Speak straight into your browser and watch the transcript appear in real time. Final results are saved to your library when you stop.
  </Card>
</CardGroup>

## Start here

<CardGroup cols={2}>
  <Card title="Dashboard" icon="grid-2" href="/en/Systems/Speech-to-Text/Dashboard">
    The Faheem home page: pick **Upload File**, **Live Transcription**, or **Record Audio**.
  </Card>

  <Card title="Upload Audio" icon="cloud-arrow-up" href="/en/Systems/Speech-to-Text/Upload-Audio">
    Drop in MP3, WAV, M4A, MP4, OGG, WEBM, or FLAC. Pick language, enable Detect Speakers, hit **Start transcribing**.
  </Card>

  <Card title="Live Transcribe" icon="microphone-lines" href="/en/Systems/Speech-to-Text/Live-Transcribe">
    Browser-based real-time transcription. Tap **Start**, speak, tap **Stop** to save.
  </Card>

  <Card title="File Library" icon="folder-open" href="/en/Systems/Speech-to-Text/File-Library">
    Every transcript you've ever made. Pin, rename, export, or delete from the 3-dot menu.
  </Card>
</CardGroup>

## Pages in this system

<CardGroup cols={2}>
  <Card title="Dashboard" icon="grid-2" href="/en/Systems/Speech-to-Text/Dashboard">
    Hero with feature specs, **How It Works** strip, three method cards, supported audio formats strip, **Pro Tips**.
  </Card>

  <Card title="Upload Audio" icon="cloud-arrow-up" href="/en/Systems/Speech-to-Text/Upload-Audio">
    600 px modal with the dropzone, file size and type rules, language selector, and the **Detect Speakers** + **Include subtitles** switches.
  </Card>

  <Card title="Live Transcribe" icon="microphone-lines" href="/en/Systems/Speech-to-Text/Live-Transcribe">
    700 px modal with a live-updating transcript area, copy-to-clipboard, and a one-click save when you stop.
  </Card>

  <Card title="File Library" icon="folder-open" href="/en/Systems/Speech-to-Text/File-Library">
    **Latest from the library** table with the 3-dot menu (**Quick export**, **Pin**, **Add to file**, **Rename**, **Delete**), a rename modal, and a delete confirmation.
  </Card>

  <Card title="Transcript Editor" icon="pen-to-square" href="/en/Systems/Speech-to-Text/Transcript-Editor">
    Two-panel editor with editable speaker turns, an audio timeline, and a right panel with **Transcribe Settings**, **Summary** / **Intent** tabs, and **Cue Properties**.
  </Card>
</CardGroup>

## Supported audio formats

| Format     | Extensions     | Max length |
| ---------- | -------------- | ---------- |
| MP3        | `.mp3`         | 5 hours    |
| WAV        | `.wav`         | 5 hours    |
| MP4 audio  | `.m4a`, `.mp4` | 5 hours    |
| Ogg Vorbis | `.ogg`         | 5 hours    |
| WebM audio | `.webm`        | 5 hours    |
| FLAC       | `.flac`        | 5 hours    |

Files must be no larger than 500 MB.

## Common questions

<AccordionGroup>
  <Accordion title="Which language should I pick when I upload?">
    Pick the language your audio is mostly in. If your speakers switch between Arabic and English mid-sentence, choose the dominant language and Faheem will still transcribe the second one in most cases. For mixed-language jobs, **Arabic** usually gives the best results because the engine is Arabic-first.
  </Accordion>

  <Accordion title="Why does my transcript only have one speaker?">
    The **Detect Speakers** switch on the [Upload Audio](/en/Systems/Speech-to-Text/Upload-Audio) modal needs to be on for diarization to run. If it was off, re-upload the file with it on. You can also rename, merge, and split speakers manually in the [Transcript Editor](/en/Systems/Speech-to-Text/Transcript-Editor).
  </Accordion>

  <Accordion title="Can I transcribe a video file?">
    Yes. Drop the `.mp4` straight into the upload modal, Faheem extracts the audio for you. The 5-hour limit applies to the duration, not the file size on disk.
  </Accordion>

  <Accordion title="My recording is longer than 5 hours.">
    Split it into smaller files (any audio editor will do). After each part is transcribed, you can copy the text together yourself.
  </Accordion>

  <Accordion title="Where do my transcripts go after I stop the live transcription?">
    Into your [File Library](/en/Systems/Speech-to-Text/File-Library). Live recordings are auto-named **Live Transcription HH:MM** with the start time. You can rename them from the 3-dot menu.
  </Accordion>

  <Accordion title="Can other people on my team see my transcripts?">
    No. Speech-to-Text files are scoped to your account. To share a transcript with a teammate, export it to TXT, DOCX, SRT, or VTT and send the file directly.
  </Accordion>
</AccordionGroup>