10 hands-on Jev projects, from drone control to a trading bot

Hoodie @HoodieFade_Originally posted 2026.10.05

Open original on X

I don't understand why nobody is using these tools yet. Jev is at the peak of its popularity right now. If you already have access to the Jev API, the best way to understand what it can actually do is to run a few real experiments instead of reading another tutorial. Here are 10 projects, ranked from the most impressive to the more niche: 1. jev-drone Use Jev to control a drone at a high level. The flight controller still handles stability and safety, while Jev makes decisions like when to take off, brake, and navigate around obstacles. https://github.com/RomanSlack/jev-drone 2. agent-desktop Let Jev control your desktop. It reads the system accessibility tree and figures out which button, menu, or input field to use next. https://github.com/lahfir/agent-desktop 3. jev-trader An experiment with high-frequency market making on the Monad testnet. Jev chooses the next buy or sell based on spreads and trade direction, with around 81 ms model latency. https://github.com/jarrodwatts/jev-trader 4. OneVOneJev A 1v1 FPS running in the browser. On every decision tick, Jev chooses movement, camera angle, aiming, shooting, and jumping. https://github.com/emrickgarrett/OneVOneJev 5. Canny A safeguard against Coding Agents that are too quick to say "done." Canny looks at tool outputs, code diffs, and test results, then decides whether the completion claim is actually reliable. https://github.com/qkal/Canny 6. Prism Jev doesn't place trades directly here. Instead, it identifies market states like toxic flow, market pressure, mean reversion, and others, then hands control back to the existing strategy. https://github.com/irfndi/prism-liquidity-agent 7. typesafe-mario Make Jev play Super Mario. No screenshots. It reads the emulator's structured RAM state and decides when to run, jump, or dodge. https://github.com/fhshaik/typesafe-mario 8. jev-curate Use Jev to filter training data. It evaluates JSONL / Parquet data for quality, relevance, and risk, then decides what should make it into the next training round. https://github.com/AkashPriyadarshii/jev-curate 9. neo4jev Put Jev inside a knowledge graph. At each node, it chooses the most promising edge for the next step and keeps following the graph. https://github.com/jexp/neo4jev 10. killmyidea Give Jev a startup idea. It evaluates it from multiple angles and ends with one of three verdicts: KILL, FIX, or SHIP. https://github.com/monteduro/killmyidea If you have API access, these projects are one of the fastest ways to see where Jev can actually be useful. Save these examples. Then check out the full Jev setup below ↓

This post quotes another post: x.com/hoodiefade_/status/2106362833189339348