Sagar here from Joyspace AI. We've built a video understanding model that works alongside transcripts to deeply comprehend people, objects, themes, personas, and the core message of a video.
Using this model as the foundation, we’ve launched Joyspace AI Clips. Joyspace AI Clips automatically creates short, shareable clips from longer videos like webinars, podcasts, talks, sales calls, Zoom recordings, etc. Our initial testing shows higher accuracy and capturing all key moments from long videos.
Right now, our model performs best when there’s a human presence in at least part of the video. It doesn't need to be every frame, but having a person in the video boosts accuracy significantly. We will update our model for other categories such as gaming videos or faceless videos. To be clear, they still work today but accuracy may not be higher.
Joyspace AI Clips works quite well with presentation-style content. From what we’ve seen, horizontal videos tend to perform better on social when there's a presentation involved, while vertical videos work better for people-first, face-to-camera formats.
All the AI features are currently free while we figure out pricing. We want to keep video transformation as accessible and affordable as possible.
We’d love for you to try it out and share any feedback – technical, product, or even just impressions. It would really help us improve.
Try it here: https://app.joyspace.ai Website: https://www.joyspace.ai
Please try it out and give us feedback. You can submit feedback and feature requests at https://www.joyspace.ai/feedback
Looking forward to hearing your thoughts.