Hey Indie Hackers 👋
I’ve been working on something new that started from a very personal pain point, and I’d love your feedback before launching it on Product Hunt.
I record tons of short vertical videos — trips, life moments, nature, friends.
At some point I had over 1500 clips sitting on my phone, and I wanted to turn them into real photo albums.
But extracting good frames manually was: slow, inconsistent , boring and impossible at scale
Scrubbing a timeline isn’t fun when you have thousands of clips.
So I built a tool I wished already existed.
BestFrame AI automatically extracts the best, most aesthetic frames from any video — in seconds.
Upload a video → AI analyzes → Get your best shots.
No timeline scrubbing. No manual work.

BestFrame AI also supports prompt-guided extraction using CLIP by OpenAI.
Meaning you can ask for:
landscape
portrait
animals
sport action
…and the AI will keep only the frames that match your prompt.
👉 If you don’t provide a prompt, it extracts the most visually aesthetic frames by default.
BestFrame AI uses:
- CLIP (OpenAI) → text–image semantic alignment
- Vision-based scoring → sharpness, composition, clarity
- Similarity filtering → avoids duplicates
Works for short clips and long videos.
Upload a video and get your best frames instantly.
I’d love your thoughts — performance, UX, features you think would help, or things that feel missing.
Happy to answer any technical questions in the comments 🙏
WOW i like it and the introduction was very good!!
Thank's, feel free to try the service and give me of your feedbacks!
Have you thought about selling BestFrame AI?