Transcribela

Features

A complete transcription workspace that runs on your own computer.

Updated November 13, 2026

Subtitles that follow professional rules

Most tools cut captions wherever the AI paused. The subtitle studio wraps every line to your limit, keeps two-line captions balanced and shows the reading speed of each one, so you can fix fast captions before you publish.

  • Characters per line, lines, and minimum and maximum duration
  • Reading speed shown for every subtitle
  • SRT and WebVTT downloads that follow your settings
Thanks for making time. Walk me through how it started.
14.6 cps
It started as a weekend experiment, honestly.
13.2 cps
We had three hundred hours of tape and no budget at all.
47.5 cps

Quotes and audio clips in two clicks

Pick the lines you need. Copy them as quotes with timestamps for your article, or save the exact audio as a WAV clip for social posts and audiograms.

  • Copy quotes with their timestamps
  • Download the selected audio as a clip
  • Works for interviews, podcasts and research

Copy quotesDownload audio clip

Chapters and show notes from your own words

The Summary tab picks key sentences, topics and YouTube-ready chapters straight from the transcript. It runs on your device like everything else.

  • Chapters that start at 0:00, ready for YouTube
  • Key points and topic list
  • A show notes draft to edit and publish
00:00 Introduction, project, budget
04:12 Laptop, model, results
09:40 Interviews, archive, search
15:05 Lessons, workflow, privacy

A private library that remembers your work

Every transcript is saved in your browser. Search across all of them, reopen any one, share a read-only link or back everything up to a file.

  • Full-text search across every transcript
  • Share links that carry the text inside the link itself
  • Backup and restore between computers
  • Founder interview.m4a42 min, Nov 3: ...three hundred hours of tape and no budget at all...
  • Board meeting October.mp41.3 hr, Oct 28: ...the budget for next quarter...

Everything else

On-device processingWhisper runs in your browser through WebGPU or WebAssembly. Recordings are never uploaded.99 languagesTranscribe in the spoken language or translate into English in one pass.Four accuracy levelsFast (Whisper Tiny), Balanced (Base), Accurate (Small) and Best (Large v3 Turbo).Clean up audioFilters rumble and hiss and evens out volume before transcription.Batch transcriptionAdd many files at once. They are processed one after another.Interactive editorClick text to edit, click a time to play, find and replace across the transcript.Speaker labelsLabel up to four speakers and name them. Labels appear in every export.My correctionsTeach it names and terms once. Corrections apply to every new transcript.Real Word documentsDownload a genuine .docx with timestamps and speaker names, plus PDF, TXT, Markdown, CSV and JSON.Library, search and backupSearch every transcript, reopen any of them and back up to a file.Share linksSend a read-only transcript. The text travels inside the link; nothing is stored on a server.Installable appInstall it on desktop or Android. On Android, share audio straight into it.Works offlineOnce a model is saved, transcription works without an internet connection.21 free toolsSubtitle converters, timing fixers, a quality checker and calculators.

Your next transcript is a drag and drop away

No account, no upload, no limits.

Transcribe a file
Summarize this page with
ChatGPTClaudeGeminiPerplexityGrok