Updated
What is context engineering?
Context engineering is the practice of controlling what an AI coding agent reads — which documents, how current they are, and how much loads per query — so it works from the right, current information instead of whatever it can scrape together. It manages the agent's inputs, not just its prompt.
Prompt engineering shapes the instruction; context engineering shapes the information the agent retrieves to follow it. As a repo and its docs grow, the second matters more — a perfect prompt over stale or scattered context still produces a wrong answer. The work is choosing the one current doc the agent should read, not handing it the whole repo to sift.
trovex is context engineering as infrastructure: it serves one canonical doc per query, at the section level, with a freshness marker, so an agent's context stays current and minimal by construction — about 60% fewer tokens per lookup. It runs locally (SQLite + ONNX, no cloud or keys).
Engineer your agents' context.
One current answer per query, with a freshness marker, in about a minute.
uv tool install git+https://github.com/TsukumoHQ/trovex
Open source. No cloud, no API keys. Your docs never leave your machine.