Updated
What is a context window?
A context window is the maximum amount of text — measured in tokens — a model can consider at once: the prompt, any retrieved docs, and its own output combined. When it fills, older content is dropped or degraded, so a bigger window raises the ceiling but does not make rereading docs cheaper.
The window is a size limit, not an efficiency gain. Every token you load into it — including the same docs an agent rereads each session to find which is current — is paid for, every time, whether the window is 8K or 2M. A larger window lets you stuff in more; it does not stop the agent from rereading or paying for stale and duplicate copies.
trovex works the other lever: instead of filling a bigger window, it serves the one canonical doc that answers a query, at the section level, with a freshness marker — so the agent reads less and reads current, about 60% fewer tokens per lookup.
Read less, read current.
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.