About
What is Spoken Wiki?
Spoken Wiki is a project that aims to create videos with audio and image from any Wikipedia article in english without human intervention.
Why?
Having the articles in video format have a number of use cases:
- Making articles more accessible to people with visual impairments or reading difficulties.
- Videos or audio can be consumed in the background when a person is doing other tasks.
- Can be shared in the same way as a podcast to wider difusion.
How?
To select which videos to process this project uses Wikimedia APIs to get trending articles from Wikipedia.
The video pipeline then converts each article into a video automatically:
- The mediawikiapi library downloads the article text and images from Wikipedia.
- Kokoro converts the text into spoken audio while the images are prepared for each section.
- FFmpeg assembles the audio and images into the final video.
More details on the implementation will be added to the blog at a later date.
Versions?
The version shown next to an article indicates which version of the Spoken Wiki video-generation pipeline was used to create its video.
- v1: The first generation of the video pipeline, using Mozilla TTS
- v2: Same as the first version but uses Kokoro TTS for audio synthesis
- v2.1: Added 1s delay at the end of each section: title and content