Cloakroom is an open-source privacy tool that masks secrets and sensitive data from AI agents. It swaps real values for placeholders, then swaps them back to the originals locally.
I was trying to build a full on local LLM setup on my machine, but the local models were not up to the mark (at least, not yet). I ended up building a hybrid setup, with the cloud for planning and local models for execution. I realized I needed a privacy layer so cloud models could not access sensitive data, and I built Cloakroom.
It runs on your machine, it’s cross-platform, and it works with Claude Code, Cursor, Gemini CLI, and Codex CLI.
Comments URL: [https://news.ycombinator.com/item?id=50446007](https://news.ycombinator.com/item?id=50446007)
Points: 1
source & further reading
github.com — original article
Show HN: Incremental Delta-Bar-Delta (Sutton's 1992 Rule) in Rust
Show HN: I have trained 15 blood glucose forecasting models on my own data
Woolly, a local memory shared by all your AI tools