Dataspaces

A Dataspace is the ownership boundary in River. Every entity — every doc, task, email, agent, conversation, and event — lives inside exactly one Dataspace. Dataspaces define who can see what, who can collaborate, and where automation runs.

You can belong to many Dataspaces at once: your personal one, your team's, and any others you've been invited to as a guest or collaborator.

What lives inside a Dataspace#

Each Dataspace is a complete, self-contained world model:

• Its own entity tree (Groups, docs, tasks, mail, files…)
• Its own Spaces and Windows
• Its own Agents, Workflows, Schedules, and Skills
• Its own Event Stream (world-events:{dataspaceId})
• Its own integrations and connected accounts
• Its own members and their permissions

Nothing crosses the boundary unless you explicitly share it.

Three kinds of Dataspaces#

Personal Dataspace#

Created automatically when you sign up. Yours alone. This is where private notes, personal email, your own AI conversations, and anything you don't want a team to see all live.

Team Dataspace#

A shared workspace for your organization or team. Everyone who's a member sees the same entities, edits the same docs in real time, and shares the same Agents. New entities created inside it are visible to all members by default.

Guest Dataspace#

When someone shares a Space, Group, or entity with you from their Dataspace, you appear as a guest. You can see and act on what was shared without having access to the rest of their world.

Switching between Dataspaces#

The Dataspace switcher lives in the top-left of every Space. Switching changes the entire context: sidebar, search, AI memory, recent activity. Your personal AI thread doesn't see your team's, and vice versa — by design.

Why this matters#

Most apps blur the line between personal and team work — your inbox is private but your CRM is shared, your notes are private but your project tracker is shared, and the AI gets confused about which context it's in.

River makes the boundary explicit. The AI in a team Dataspace can reason across everyone's docs, tasks, and emails because they're all in the same world. The AI in your personal Dataspace can't, because that data is yours alone.

This is how you keep collaboration powerful and privacy intact at the same time.

Tips#

• A common pattern is one personal Dataspace + one team Dataspace per organization.
• Move entities between Dataspaces with right-click → Move to — this is a real move, not a copy.
• Need to keep certain projects extra private inside a team Dataspace? Use Group-level permissions instead of creating a new Dataspace. See Permissions.