Introducing Proof — Dan Shipper

Every launched Proof, a free, open-source document editor built for agents and humans to collaborate. Key features: live editing with multiple collaborators, comments, change tracking, and visual provenance tracking (green rail = human-written, purple rail = AI-written).

The thesis: most word processors assume a human is writing with AI at the margins. Proof inverts this for the kinds of documents agents increasingly produce — bug reports, PRDs, implementation plans, research briefs, strategy documents, and memos. Proof is agent-native (anything a human can do, an agent can do), tracks provenance visually, and requires no login.

Internal use cases at Every: Austin Tedesco uses it for campaign strategy documents, iterating with multiple agents before sharing on Slack. Shipper uses Proof documents as daily to-do lists, with his Claw R2-C2 adding tasks throughout the day. The editorial team is exploring integration into their AI writing workflow. The initial draft of the announcement post itself started in Proof.

RDCO Mapping

Worth monitoring as a potential tool for our own agent operations. The provenance tracking (human vs AI visual markers) addresses a real transparency concern in agent-generated content. The open-source, no-login approach suggests Every is building ecosystem infrastructure, not just products. Could be relevant if we need shareable agent-generated documents in our content pipeline.