Jev ecosystem scrape: routing is crowded, search and memory are open
A scrape of early Jev ecosystem data finds 19.6% of projects take 63% of stars, routing and simulation have similar project counts but a 45x gap in attention, and search, memory, and expert tasks remain open.
someone scraped the entire early Jev ecosystem and the result is a treasure map for anyone deciding what to build 19.6% of Jev projects pull 63% of all the GitHub stars. that's routing and interface agents. everyone is building there. meanwhile search & memory has 381 projects averaging 37 stars, and content & expert tasks has 387 projects at 31. the line that stopped me: routing and simulation have almost the same number of projects, 250 vs 252. one averages 364 stars, the other 8. same amount of work, 45x the attention. and Jev isn't doing the same job everywhere. in simulation it mostly picks actions (52%). in routing it picks which model or tool to call next (27%). in most categories its main job is simply judging whether something is what it claims to be. if you're building on Jev right now, the crowd is in routing. the room is in search, memory and expert tasks
This post quotes another post: “Paper maps early Jev ecosystem usage against attention”