4 ms·
Show HN: Shotty – Create TikTok/Instagram-like shorts from a given URL or text
Hey HN, this is just a fun weekend project called Shotty. It lets you create TikTok/Instagram-like shorts from a given URL or text.
Please don't take generated content seriously :)
Repo: https://github.com/n4ze3m/shotty-fun https://github.com/n4ze3m/shotty-fun
Few generated videos:
https://shotty.fun/api/v1/file/clpzn4l110003v0ntxwf1w9ac/video https://shotty.fun/api/v1/file/clpzn4l110003v0ntxwf1w9ac/vid...
https://shotty.fun/api/v1/file/clpznys9x0005v0ntvtcyovct/video https://shotty.fun/api/v1/file/clpznys9x0005v0ntvtcyovct/vid...
- error404x 3y agoAlso Video processing may take a few minutes because of my VPS. I'm welcoming feedback. Thanks :)
- error404x 3y agoNote: Currently, the application does not load PDFs or YouTube videos.
- smcleod 3y agoReally neat little project! It’d be really cool to see a demo of this using only self-hostable components (e.g. ollama, whisper, llava vision)
- error404x 3y agoThe whisper I'm using is self-hosted on the same VPS, but yeah, I will try llava vision. Thanks
- mdrzn 3y agoGreat work!
- O1111OOO 3y agoWe seem to take so much for granted these days but this fun weekend project is literally the future. This is Star Trek level of creation. Pretty soon, we'll have these kinds of creation tools 'read' a book and create a 2 hour movie with SFX. Read a comic series and create a franchise. Listen to the humming of a tune and create a complete music video. The "creator" in some cases, will continue to tweak and tweak until it becomes a masterpiece - and in doing so... become part of the creative process. In other cases - as were seeing now - we'll be bombarded with tons of low-effort spam. This was always going to be the price paid for these kinds of tools. We all know this. Still... we're going to see so many more people actively creating because they're able to do so now with these kinds of tools. Thanks for making this - it was fun testing a couple of sites!
- error404x 3y agoUpdate: Since my ElevenLabs credit ran out, I'm now using the SpeechT5_TTS model.