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

# Transcription

> Accurate speech-to-text transcription powered by advanced AI.

## Overview

SkyScribe's transcription feature converts spoken audio into written text in the same language as the original audio. This is ideal when you want an exact written record of what was said.

## How to Transcribe

Choose your preferred method to get started:

<CardGroup cols={3}>
  <Card title="Upload" icon="upload" href="/quickstart/upload">
    Upload an audio or video file from your device
  </Card>

  <Card title="Fetch" icon="link" href="/quickstart/fetch">
    Import media from YouTube, TikTok, or other public URLs
  </Card>

  <Card title="Record" icon="microphone" href="/quickstart/record">
    Record audio directly from your device
  </Card>
</CardGroup>

**When selecting your settings:**

1. Choose **Transcribe** (not Translate)
2. Select the source language or use auto-detect
3. Configure optional features (Speaker Diarization, etc.)

## Common Use Cases

<AccordionGroup>
  <Accordion title="Meeting Notes" icon="users">
    Convert meeting recordings into searchable, shareable text documents.

    **Example:** Record team meetings and automatically generate notes for attendees who couldn't join or for future reference.
  </Accordion>

  <Accordion title="Interview Transcripts" icon="comments">
    Transform interview recordings into written transcripts for analysis and quotes.

    **Example:** Transcribe research interviews, podcast episodes, or job candidate interviews for easier review and editing.
  </Accordion>

  <Accordion title="Lecture and Educational Content" icon="graduation-cap">
    Create written versions of educational videos and lectures for studying and accessibility.

    **Example:** Convert recorded lectures into study materials that students can search, highlight, and review.
  </Accordion>

  <Accordion title="Podcast and Video Content" icon="podcast">
    Generate text versions of audio and video content for SEO, accessibility, and content repurposing.

    **Example:** Transcribe podcast episodes to create blog posts, show notes, or searchable archives.
  </Accordion>

  <Accordion title="Legal and Compliance Documentation" icon="scale-balanced">
    Create accurate records of legal proceedings, depositions, or compliance calls.

    **Example:** Transcribe legal consultations, court proceedings, or compliance training sessions for official records.
  </Accordion>
</AccordionGroup>

## Key Features

### Speaker Diarization

Identify and label different speakers in multi-person conversations.

**When to use:**

* Meetings and conference calls with multiple participants
* Interviews with distinct speakers
* Panel discussions and roundtables
* Podcasts with multiple hosts

**Example output:**

```
Speaker 1: Welcome to today's meeting. Let's start with updates.
Speaker 2: I've completed the project analysis. Here are my findings.
Speaker 1: Excellent. Can you share more details?
```

### Segmentation Styles

Choose how your transcript is broken into segments:

<CardGroup cols={2}>
  <Card title="Readable" icon="book">
    **\~30 second segments** with punctuation and capitalization

    Best for documents, meeting notes, and general transcription
  </Card>

  <Card title="Subtitle" icon="closed-captioning">
    **Short, punchy segments** optimized for video captions

    Best for video captions, subtitles, and social media content
  </Card>
</CardGroup>

## Improving Transcription Quality

### 1. Use High-Quality Audio

Better audio quality leads to more accurate transcriptions:

* Clear speech with minimal background noise
* Good microphone and recording equipment
* Proper audio levels (avoid too quiet or distorted audio)
* Reduce echo and reverb when possible

### 2. Specify the Language

While auto-detect works well, manually specifying the language can improve accuracy:

* **Auto-detect** - Convenient for single-language, clear audio
* **Manual selection** - Better accuracy when you know the language, especially for mixed-language audio, specialized dialects, or consistent processing

<Tip>
  For best results, use manual language selection when you know what language is spoken in your audio. See our [Language Support guide](/guides/language-support) for details.
</Tip>

## Transcription vs Translation

Understanding when to use each feature:

<CardGroup cols={2}>
  <Card title="Use Transcription" icon="file-lines">
    When you want text in the **same language** as the audio

    **Example:** English audio → English text
  </Card>

  <Card title="Use Translation" icon="language" href="/guides/translation">
    When you want text in a **different language** from the audio

    **Example:** Spanish audio → English text

    Learn more about translation →
  </Card>
</CardGroup>

## Output Formats

Download your transcripts in multiple formats:

* **PDF** - Formatted document for sharing and printing
* **DOC** - Microsoft Word format for editing
* **TXT** - Plain text for simple use cases
* **MD** - Markdown format for documentation
* **SRT/VTT** - Subtitle formats with timestamps
* **CSV** - Structured data with segments and timestamps
* **JSON** - Programmatic access with full metadata

## Supported Languages

SkyScribe supports transcription in **99 languages**. Language detection can be automatic or manual.

<Card title="View All Languages" icon="language" href="/guides/language-support">
  See the complete list of supported languages, language detection guidance, and best practices.
</Card>

## Best Practices

1. **Test your setup** - Record a short test clip to verify audio quality before long sessions
2. **Choose the right format** - Use readable segments for documents, subtitle segments for videos
3. **Use speaker diarization for conversations** - Makes multi-person transcripts much easier to follow
4. **Review critical transcripts** - Always review transcriptions for business-critical or legal use

## Limitations

Transcription accuracy varies based on audio quality, accents, background noise, and content complexity. For detailed information:

* See [Language Support](/guides/language-support) for language-specific details and accent variations
* Review [Model documentation](/guides/whisper-model) for technical limitations and performance characteristics

<Note>
  We recommend testing with your specific content type and reviewing output for critical use cases.
</Note>

## What's Next?

You've learned about SkyScribe's transcription capabilities and features.

**Want to learn more?**

* Learn how to [edit your transcripts](/guides/transcript-editor) with AI-powered tools
* Explore [translation](/guides/translation) to convert transcripts to other languages
* Check out [language support](/guides/language-support) for details on specific languages

## Need Help?

If you have questions about transcription capabilities or quality:

* Check our [Language Support guide](/guides/language-support) for details on specific languages
* Review the [Model documentation](/guides/whisper-model) to understand the AI powering transcription
* Contact our support team at [support@sky-scribe.com](mailto:support@sky-scribe.com)
* Report transcription issues to help us improve