What it does
YobiYoba is a speech recognition service from Yobinext SAS, a company based in Orsay, France. You upload audio or video files in any format, set the language, and receive a time-coded transcript. If the language is not given, the service tries to recognise it. The result is a fully annotated document with speech and non-speech segments, speaker labels, words with time codes, confidence scores and punctuation. An editor lets you check the transcript against the audio and correct it, and you can supply a list of uncommon words or related text to improve accuracy. The site notes that no speech recognition is error free. Transcripts can be exported as PDF, DOC, RTF, CSV, SRT, VTT and other text and subtitle formats. Pricing is pay as you go: you pay only for the amount of speech transcribed, not the file length, with no minimum fee per file and credit that never expires. The site is available in English, French and German.
Capabilities
- Upload audio or video in any format and get a time-coded transcript
- Language recognition when the spoken language is not specified
- Speaker labels, word timings, confidence scores and punctuation
- Built-in editor synchronised with the audio
- Export to PDF, DOC, RTF, CSV, SRT, VTT and more
- Custom word lists and related text to improve results
When to use it
- You need a transcript of an interview, meeting or recording that you can correct yourself
- You want subtitle files in SRT or VTT from an audio or video file
- You prefer paying per minute of speech rather than a subscription
