Building an app to search across videos using AI (part 2) (stream 222)
Summary
Continued working on FrameFisher, successfully implementing video upload, delete, and search functionality using MudBlazor. Added a threshold slider to filter search results and integrated the API to search through video content using AI-powered vector search. Of course, I had to switch back to Debian from Omarchy because my dev environment kept breaking—nothing says “professional streamer” like spending half your time fixing Linux distro issues instead of actually coding.
📺 - Twitch archive - stream no.278
Replay
Project
All the code for this project is available on GitHub: frame-fisher - https://github.com/FBoucher/frame-fisher
Game Results
- @thecliptographer: 2.02
- @fboucheros: 2.31
- @theunoriginaljerk: 8.02
- @jeffs_hat_stand: 9.27
- @undefined_process: 34.98
- @a_friend_asks: 54.26
- @surlydev: 58.94
- @groversaurus: 64.22
- @my_friend_asks: 70.80
- @jtsom: 76.91
- @dj_surly: 84.75
- @lurkydev: 89.67
- @codebymistakes: 90.53
- @coppersbeard: 93.29
- @cmgbeanie: 93.54
- @fredda_the_cat: 93.87
Statistics
- 🏆Best score: @fredda_the_cat with 93.87
- 😭Biggest loser: @therealsurlybot with 21 drops and no high score
- 🍀Luckiest: @fredda_the_cat with best score 93.87 and only 21 drops
- 🎖️Super participant: @surlydev with 29 drops
Notes/ References / Snippets
- Demo Be-my-eyes https://github.com/fboucher/be-my-eyes
- Frame-fisher https://github.com/fboucher/frame-fisher
- workshop https://github.com/reka-ai/api-examples-python