Audio and video transcription
Turn audio or video files into text with speaker labels, word timestamps, and TXT/SRT/VTT export.
Speechdash transcribes audio and video files into readable, timestamped text. Transcription runs in the background: upload a file, keep using the app, and come back when it is done. Transcriptions in the sidebar is the job queue. When a job finishes, the text opens in Library.
What you get
- Full transcript split into segments with start/end times
- Speaker labels (optional): segments are tagged
SPEAKER_00,SPEAKER_01, and so on when speaker detection is enabled - Word-level timestamps for precise navigation and subtitles
- Export as plain text (TXT) or subtitles (SRT, VTT)
Supported files
Common audio and video formats work: MP3, WAV, M4A, AAC, OGG, OPUS, FLAC, MP4, MOV, WEBM, MKV, AVI, and more. Files can be up to 500 MB.
Start a transcription
Drop an audio or video file on the homepage, or open Transcriptions in the sidebar and start a new job.
Pick a language (or leave auto-detect) and toggle speaker detection if the recording has multiple voices. These options are locked once the upload starts.
Confirm. The file uploads with a progress bar, then the job is queued. You can leave the page; the queue shows live status. When the job completes, click it to open the transcript in Library.
Language detection
Leave the language on auto and Speechdash detects it from the first seconds of audio. Choosing the language explicitly helps with short files or heavy accents.
Costs and limits
Transcription bills 1 credit per 2 minutes of audio (every started minute counts). Credits are charged when the transcription completes, never for failed jobs.
| Plan | Max length per file | Simultaneous transcriptions |
|---|---|---|
| Free | 30 minutes | 1 |
| Unlimited | 10 hours | 5 |
Read and export
A finished job opens in the Library reader (same layout, voices, and playback as other documents). There is no separate transcript page: Transcriptions in the sidebar is only the job queue.
From the reader toolbar you can:
- Listen with sentence highlighting
- Export as
TXT(plain text),SRTorVTT(subtitles with timestamps and speaker tags)
In-progress and failed jobs stay on the Transcriptions queue so you can watch status.