I have a few hundred hours worth of audio I want to transcribe to text, but while Elevenlabs' Scribe v1 Speech-To-Text model is the best one in terms of quality, nuance and detail, their Web UI only offers one-at-a-time transcriptions, with slow speed.
And it turns out, if I use the Elevenlabs' API to transcribe, suddenly I get whopping 12.5x multiplier to transcribe my audio in terms of credits. That means on the Starter plan subscription I normally get 1 hour per month via their website, meanwhile via the API I'd get 12 hours and 30 minutes!
So I built this. Run the exe, choose the folder with my audio files in it, press Start and it will do the whole damn folder automatically, at turbo speed, while spending my subscription credits at a way cheaper, 12.5 times efficient rate. Massive money saved.
I made each transcription as useful as possible and made it automatically output 3 files per each audio; a TXT file for us humans, an SRT file for video subtitles, a JSON file with detailed breakdown of each word and audio event that can be heard in the file (including stuff like finger snapping, footsteps, sound effects, etc).
To celebrate the launch, you can get it 60% off between days 1.8.2025 - 15.8.2025 at:
https://itch.io/s/157701/elevenlabs-batch-transcriber-early-bird-discount
I build this with Claude Pro with a huge ass engineered prompt I made that took me several hours, plus more hours to tweak the GUI, output and other sensibilities of this program. Pretty happy how it came out.
Top comments (0)