Skip to content

v3.0.0

Latest
Latest
Compare
Choose a tag to compare
@Martouta Martouta released this 08 Apr 15:14
· 135 commits to main since this release
v3.0.0
68c7e37
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Verified
Learn about vigilant mode.

The release has been generated with make version=v3.0.0 deploy_production.

The main differences are:

  • ⚠️‼️🚨 BREAKING CHANGE: Save transcripts/subtitles in SRT format in the file, and in a similar format in MongoDB to finally include the timestamps as well.
  • Add support for multiple Speech-to-Text AI platforms: Google, OpenAI Whisper, AssemblyAI, Gladia, and Microsoft Azure.
  • Allow running resources with local paths.
  • Support also webm and m4a formats.
  • Update Python from 3.10 and 3.11
  • Do not save subtitles if the whole thing is empty
  • For YouTube, allow downloading the captions for any specific resource in any specific language if available. Allow only trying the captions and do nothing if it fails, or try downloading the captions and if it fails, then go for AI, or even just go for AI directly.
  • Save the audio locally differently to use less memory (which is noticeable for large files).
  • Update dependencies. Especially for security.

README.md and CHANGELOG.md updated 😉

Assets 2
Loading