October 2026

We are building one place for everything people and their agents make.

Code has had a place like this for decades. Every developer and every tool pushes to the same host, and that is what lets a team work on one thing together. The rest of knowledge work has nothing of the kind. Memos, decks, sheets, plans and findings end up in chat windows, shared drives and inboxes. Every agent you use keeps its own copy of the context, and none of them can see what the others did.

read write labs is building that place. It is a hosting platform for all knowledge work and everyone who does it. It holds whatever a project produces, whether that is code, a doc, a deck, a sheet or a skill. There is version control underneath, so nothing gets lost and you can see who changed what. Most of the time you won't think about the history at all. You'll just open the project.

Once the work lives in one place, a team and all its agents can work on the same project. The agents don't have to match. One person can use Claude Code on Anthropic's models and another Codex on OpenAI's. A third can run an open model on a server in the office. All of their work lands in the same project. We built it for that mix, and we are building four products on top of it.

Projects is the first, and the one to start with. It is where a team and its agents work together on one thing. Because we host the work, Projects sees every change as it lands and shows it to you.

Connectors lets you connect a tool once, say Drive or Slack or your calendar, and every agent you use can reach it from then on. You don't set it up again for each new agent.

Access lets you decide who sees which files. Safety, security and compliance matter as much as keeping everything in one place, so we built access in from the start. The whole team can read the plan while only the two colleagues who need to can open the financials. The rest of the team doesn't see that folder at all. That protects the most sensitive work without giving up one place for everything. We can do that because we host the files, and a shared folder can't.

Ambient Intelligence is where you tell us what to watch for. Say "keep this project on track and flag divergence early," and it reads along as the work changes. It speaks up when two people disagree, when the strategy has gone unclear, or when tasks drift apart. It can do that because everything it needs to read is in one place.

Every one of these needs the platform under it, which is why we are building the platform first.

Knowledge work has been read-only. We are making it read-write.

read write labs, inc