Jev Lens: a single PowerShell script turns folders into AI-ranked semantic search
dfinke · x · 2026-09-22
Developer dfinke open-sourced Jev Lens, a compact Windows demo that turns any folder into a semantic search surface: describe what you need in plain English, and the AI (Jev) makes typed, probability-backed micro-decisions over a bounded set of local files, with deterministic PowerShell ranking turning those into results.
- Intentionally minimal: one PowerShell script, one WPF view, one API request, and a local fallback — no browser, npm install, agent framework, or admin rights.
- Requires Windows and PowerShell 7.2+; set a TYPESAFEAPIKEY and run one command, with the current directory as default scan target.
- Code available at GitHub (dfinke/jev-experiments), with a YouTube demo.
Related event: Jev Lens Brings Semantic File Search to PowerShell(2 posts)→
More from coding & agent
- PE fund rolled coding agents across portfolio: only 1 of 5 workflows worked — alex_verem · 2026-09-22
- capy beats a major coding harness at half the cost and half the time, eval finds — garrytan · 2026-09-22
- XGrammar-2 ships strict tool calling for agents, adopted by xAI, DeepSeek and vLLM — vllm_project · 2026-09-22
- MechFaber: Claude Code designs a 99-part quadruped with firmware co-simulated in Renode and MuJoCo — SpeedyBrowser45 · 2026-09-22
- Exa MCP hits 5,000 GitHub stars as AI agents flock to its search integration — TheIshanGoswami · 2026-09-22
- 670,000 agent skills, no trust layer: bot scan finds 69% never reliably fire — markjeffrey · 2026-09-22