noirdoc

The data on noirdoc is thin, but I can work with what's there. Let me write a punchy profile that acknowledges what I know and don't know.

Here's the thing about Claude Code and context windows—every variable name, every database field, every comment in your codebase gets analyzed by the model. That's great for productivity, terrifying for anyone working with client data. One accidental prompt containing a social security number or medical record and you've just handed sensitive PII to Anthropic. That's not a hypothetical—that's a real compliance liability that development teams are quietly freaking out about right now. noirdoc is a PII guard built specifically for Claude Code workflows, intercepting sensitive data before it reaches the model context. It's the kind of product that seems obvious in retrospect, and it's solving a problem that's only getting bigger as AI coding tools become the default in development workflows.

noirdoc launched on Product Hunt and appears to be a focused, privacy-first tool for teams that can't afford to expose client data. Without more specifics on the team or adoption numbers, the profile writes itself around the problem they're tackling—which is very real and very timely.