SkillSwap

SkillSwap is for people with a side hustle, no budget, and one missing skill. You list what you can teach and what you want to learn, and the app finds someone to trade with. The simplest match is a direct swap: you teach what they want, they teach what you want. When no two people line up neatly, it finds a loop instead, where A teaches B, B teaches C, and C teaches A, and everyone still gets what they need.

When neither a swap nor a loop exists, time credits fill the gap. Everyone starts with 2 credits, learners pay hours to a teacher, and sessions can be booked, completed and reviewed. Matching is not just string comparison: a small local model understands that “figma” and “ui design” mean nearly the same thing, so people who describe the same skill differently still find each other, and if no model is running the app falls back to exact matching and keeps working.

Everything runs locally with no API bill and no account on someone else's server: your skill profile, which is really a list of what you cannot do yet, never leaves your own machine. The interface follows Mac conventions with dark mode included, and a demo login is seeded so you can try the whole flow from sign-up to booking to review.

Honest limits: there is no hosted demo, so you run it locally with Python and optionally Ollama; the live deployment notes in the repository describe a Render setup but no public demo link is given. The similarity threshold for fuzzy matches is a starting point the author has not yet tuned on real users, and real scheduling, in-app chat and new-loop notifications are still on the to-do list.