News · New this week
3 GitHub repos with real AI projects, and the code
Most AI project lists are only links. These three ship the code, and one of them has forks you need to avoid.
You searched "AI project ideas" and got a hundred listicles. Every one of them handed you a title and a one-line description, and none of them handed you anything you could run. The ideas were never the hard part.
These three GitHub repos are different because the code comes with them.
awesome-llm-apps
github.com/Shubhamsaboo/awesome-llm-apps
Its own title says 100+ AI agents, agent skills and RAG apps, all free and open source. Each one is a folder you can clone and run, not a link to someone's blog post about it.
Start here if you want something working on your machine tonight rather than a reading list.
500 AI, ML, DL, CV and NLP projects
Five hundred projects with code, sorted by topic, so you can go straight to computer vision or NLP instead of scrolling past everything else.
One warning on this one. It has dozens of near-identical forks sitting at 10 to 80 stars, with the same repo name. The real one is under ashishpatel26. If you search the name instead of using the link, check the owner before you clone.
ai-engineering-hub
github.com/patchy631/ai-engineering-hub
In-depth builds rather than a long list. Real LLM, RAG and agent applications, walked through properly, organised by difficulty so you can pick something at your level instead of bouncing off the first repo you open.
On the numbers
The counts above are what each repo claims in its own title. I have not counted the folders myself, and you should treat them as the maintainer's figure rather than a verified total.
Star counts move too. These were checked on 21 October 2026, so they will have drifted by the time you read this.
What to actually do with them
Pick one project. Clone it. Get it running before you change anything.
The reason most project lists go nowhere is that reading them feels like progress. Running one thing, badly, teaches you more than saving thirty.
