Clips Kitty local-first AI clipping

Roadmap

Where Clips Kitty is going, and what already works. Anything listed as shipped is in a release you can install today, not a plan.

Now: stable releases

Clips Kitty is a stable release. Testing on machines that are not the developer's has been worth it from the start: the first release could not finish a clip on any source, and nobody could have known without installing it somewhere clean.

  • Shipped. One Windows installer carrying the app, the engine, FFmpeg, the AI runtime and the tracking and transcription models. Releases with notes, a changelog and a known-issues list. Bug reports and feature requests, including an in-app button that needs no GitHub account. Discussions. Security scanning. A container image for working on the engine. This website.
  • Publishing. A clip, or a whole batch, posted to YouTube, TikTok, Instagram, Facebook, X, Threads, LinkedIn, Pinterest and Bluesky at once, with a posts-a-day schedule so a batch does not land in one burst. It goes through WoopSocial or Upload-Post on your own account, and it is off until you turn it on.
  • Sports: Soccer and Basketball. A match's goals, saves and cards, or a basketball game's dunks, threes, blocks and game winners, each as its own clip, framed on the ball. More on sports clips.
  • Being worked on. Code signing, so Windows stops warning that the installer is unsigned. Screenshots and a demo video. Progress reporting for the scoring stage and for update downloads, so working cannot be mistaken for hung. Automated Windows builds.
  • Plugins. Plugins that find, understand and rate the moments in a video are built, and so are plugins that suggest edits for the clips, which wait for you in the editor, with a Python SDK for writing them. Clips Kitty 2.1.0 is the first release that runs plugins, and it comes with SDK 1.3.0. For developers.

Affiliate link - Clips Kitty may earn a commission if you sign up through it, at no extra cost to you.

Next: community

Making it worth someone's time to contribute: good first issues with enough context to pick up cold, contributor recognition, example workflows, community showcases, performance benchmarks, and voting on feature requests.

Contributions are wanted in features, bug fixes, documentation, translations, AI improvements and performance work.

Later

None of this holds up a release.

  • More sports. Soccer and Basketball are in. Cricket is next.
  • Text updates when things finish. A text message when a video's clips are ready, when they're posted, and when something fails.
  • Reaction videos. These need the app to understand the video being reacted to, not just the words spoken over it.
  • An Android companion app, Vertical Live only, so it stays lightweight: Twitch, Kick and local files, within Google Play's policies.
  • Creator analytics. Plus more models and platforms where they earn their place.

The long-term idea

Clips Kitty is local-first, and that word carries the whole design: your footage, your workflows and your data stay on your machine. The app gets better as your hardware does, rather than as someone's subscription tier does.

Consumer AI hardware is improving quickly. The architecture is kept modular so newer models and faster machines can be adopted without a redesign. As hardware like the NVIDIA RTX Spark and future AI-accelerated GPUs arrives, the same application should run larger local models, infer faster, analyse video better and automate more.

The goal is an open-source creator platform that becomes more capable over time because local AI hardware does, not one that becomes more expensive.