# ScribeOS Web

## Core Features

### Batch Transcription

Batch transcription processes pre-recorded audio or video files into structured, timestamped text.

#### Supported Formats

Audio

* MP3
* WAV
* M4A
* FLAC
* OGG
* WEBM

Video

* MP4
* MOV
* AVI
* MKV
* WEBM

#### File Limits

* Audio: Maximum 1GB
* Video: Maximum 2GB
* Supported sample rates: 8kHz–48kHz
* Mono and stereo supported (stereo is automatically downmixed)

#### Processing Engine

Batch transcription is powered by Whisper v3 Turbo from OpenAI.

#### Processing Lifecycle

Each batch session follows a defined state machine:

1. Pending — job created
2. Queued — waiting for processing
3. Processing — actively transcribing
4. Completed — transcript ready
5. Failed — error occurred (retry available)

#### Output Format

Transcripts include:

* Timestamped segments
* Automatic punctuation
* Capitalization
* Language detection metadata

#### Performance Expectations

* \~78x real-time processing speed
* 6-minute audio typically completes in under 20 seconds
* Accuracy varies by audio quality:
* Clear audio: up to 99%+
* Moderate noise: 95–98%
* Heavy accents or multiple speakers: 85–95%

### Live Transcription

Live transcription converts streaming microphone input into real-time text.

#### Streaming Engine

Powered by Nova-3 from Deepgram.

#### Technical Characteristics

* WebSocket-based real-time connection
* Audio chunks processed approximately every 100ms
* Interim and final transcripts supported
* Automatic cutoff at 1 hour

#### Transcript Behavior

* Interim transcript: temporary, updates continuously
* Final transcript: stabilized and stored permanently
* Entire session saved automatically upon completion

#### Session Limits

* Maximum duration: 60 minutes
* Credits rounded up to nearest minute
* Session auto-terminates at limit

#### Common Use Cases

* Live meetings
* Real-time captioning
* Customer support calls
* Medical dictation
* Live podcast recording

### Translation

Translation converts completed transcripts into another language.

#### Translation Engine

Powered by GPT-5-mini from OpenAI.

#### Capabilities

* 40+ supported languages
* Context-aware translation (not word-by-word)
* Timestamp preservation
* Structural formatting retained

#### Billing Model

* 1 credit per \~3,000 tokens
* Tokens calculated from source transcript length

#### Quality Considerations

* Idiomatic expressions are adapted contextually
* Technical terminology handled probabilistically
* Human review recommended for legal or medical documents

### Session Management

All transcriptions are stored as sessions within the user dashboard.

#### Status States

* Pending
* Queued
* Processing
* Completed
* Failed
* Streaming (live in progress)

#### Organizational Tools

* Custom session titles
* Bookmarking (starred sessions)
* Search by title
* Filter by status
* Sort by date or alphabetical order
* Bulk selection actions

#### Session Persistence

* Sessions remain stored until deleted
* Export available in supported formats


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://scribeos.gitbook.io/scribeos/scribeos-web.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.