Engineering deep-dives, product updates, and field notes from building consent infrastructure for the agentic web.
AI coding agents now write meaningful parts of the web. We built consent infrastructure for the engineers — and the agents — that ship them. Here's the thinking behind /.well-known/consent.json, the MCP toolset, and why we started there instead of with a banner.
How we shipped a sub-3KB loader that runs before any tracker and passes Core Web Vitals on every site we tested.
Six layouts, full theming, headless API. The fastest way to ship a compliant banner without writing CSS.
A machine-readable consent policy isn't a feature. It's the only honest way to let AI agents respect user choices.
Brazil's privacy law has unique requirements. Here's how ConsentLayer handles them automatically by visitor IP.
Every payload is HMAC-SHA256 signed and timestamped. Here's how we built the verifier and shipped snippets for every framework.
Tell Cursor or Claude Code to "add ConsentLayer." One natural-language step from empty repo to live consent.
A platform team's story of swapping their entire CMP stack including category maps, theme rules, and audit history.
Append-only isn't enough for regulators. Here's why every change in ConsentLayer is cryptographically chained and verifiable.
We tested every major operator-style agent against 50 sites. Most flat-out ignore CMPs. A few honor signed policies. Here's the data.