XERJ's CRAZY Local Search for AI Agents: I Tested It, It Broke, Then It Got AWESOME

XERJ is an open-source, local-first search engine and MCP server for AI coding agents. I installed it on my own Windows laptop, indexed real code, and report what worked and what broke. Repo: https://github.com/xerj-org/xerj (Apache-2.0). Version tested: v1.0.0-rc.98. Not tested: the README's 2.7x fewer-tokens study (their claim, not my result). Chapters: 0:00 The failure (cold open) 0:30 What XERJ is 1:13 The claim I did not test 1:48 Install, local first 2:21 Whole-repo index: the failure 3:16 Engine-folder index: the working run 3:45 Search quality and latency 5:27 Reference coding mini-test 6:28 Elasticsearch-compatible API 7:09 Agent memory 7:51 MCP, xerj init and auto field reports 9:14 Verdict 9:51 RunPod and your turn If you are looking for cheap and reliable GPUs, I use RunPod: https://get.runpod.io/pe48 Get GPUs Runpod: https://get.runpod.io/pe48 Get Hostinger: https://www.hostg.xyz/SHJiu Coupon Code "PROMPT" 10% discount 🔗 Connect with me: Email → [email protected] Ko-fi → https://ko-fi.com/promptengineer Patreon → https://www.patreon.com/PromptEngineer975 Book a Call → https://calendly.com/prompt-engineer48/call GitHub → https://github.com/PromptEngineer48 Twitter/X → https://twitter.com/prompt48 Website → https://promptengineer48.com/