Spotify Wrapped for your commits

gm legends, happy Friday.

GitStory turns your commits into something you can actually watch, KnockKnock keeps video calls with your inner circle short and intentional, and SAM Audio from Meta lets you pull specific sounds out of a messy recording like it’s no big deal.

GitStory turns your GitHub repos into short, watchable timelines instead of static graphs and green boxes. It maps how a project changed over time so you can show the real work behind it, not just the final screenshot.

🔥 Our Take: Dev work is usually invisible unless someone is willing to dig through commit history with you. A quick visual replay is a much easier way to show how you think, what you ship, and how messy the middle really was. Feels more like showing your process, less like begging someone to read diffs.

Mona Truong opened a soft, simple thread asking everyone to share three things they’re grateful for every day. Not a once-a-year November list, but the small, repeatable things that actually keep you grounded when everything else is loud. It’s basically an open invite to pause for a second, name what’s keeping you sane, and maybe borrow a few ideas from what other people write.

KnockKnock is a tiny video-calling app for the people who actually matter to you. One-click Mac calls, no links, no waiting rooms, just floating windows that stay out of the way and a hard five-minute limit so chats never turn into meetings. Themes wrap your video in custom rings and images so every call feels personal instead of corporate.

🔥 Our Take: Video apps act like every call is a meeting. This goes the other way with strict time limits and lightweight UI so you can ping a friend or teammate, say the thing, share a moment, and get back to your day. It makes spontaneous calls feel acceptable again instead of like a calendar negotiation.

SAM Audio is Meta’s new model that lets you pull a single sound out of a messy mix with a prompt. You can type what you want to isolate, click on an object in a video, or mark a time span, and it separates that voice, instrument, or noise without training a custom model for each case. It’s open source, runs in the Segment Anything Playground, and comes in multiple sizes you can download for your own workflows.

🔥 Our Take: This goes straight at one of the most annoying parts of editing: trying to rescue one sound from a chaotic recording. Instead of hunting through plug-ins and hoping a preset works, you just tell it what you want and get a clean stem back. Creators are going to love this, and so are anyone doing podcasts, film, TikToks, or research with real-world audio that used to be basically unfixable.