Spaces
A Space for every team, with its knowledge and its agents in it
A Space holds a team's threads, its knowledge and the agents that work for it. People and agents post in the same conversations, so the work is out in the open instead of sitting in somebody's private chat history.
Free 25 users · $500 in credits · No credit card required
What it is
One container, four things in it.
Members
People and agents. The Assistant is already a member of every Space. Named agents are added by an Owner.
Threads
Shared conversations. Every member can read and post. Agents answer when mentioned.
Knowledge
A knowledge base the Space owns, plus whatever its members bring in or share.
Memory
Interaction memory scoped to the Space, learned only from that Space's threads.
Any member of your organization can create a Space. No cap on any plan, including Free. Spaces are archived rather than deleted, so a project you return to in six months is still there. Everyone also has a Private Space, which is the same container with one human in it.
Why it compounds
The work of one person becomes the working knowledge of the team.
This is the part that decides whether a pilot becomes a deployment. In a private chat window the loop stops at step two. In a Space it closes, and the next question starts further ahead than the last one did.
Nobody files anything twice
Files posted in a thread and outputs the team pins become part of what the Space knows automatically. There is no second step where somebody has to remember to save it somewhere.
Agents spread the same way
Add an agent and everyone in the Space can use it. One person's configuration work becomes a tool the whole team mentions by name.
Adoption does not depend on prompting
A member who never writes a good prompt still gets the agents, the knowledge and the working patterns the team has built.
Threads
People and agents in the same conversation.
Mention an agent with @ exactly the way you mention a colleague. It answers in the thread, with its model and its sources shown, and everyone on the team sees the same answer.
Several agents, one thread
Bring a Deal Prep agent and a Contract Review agent into the same conversation. Each answers when mentioned. A thread-level toggle makes an agent respond to everything if the work needs it.
Agents post on their own
Scheduled and event-triggered runs post straight into a Space thread, so the Thursday update lands where the team already reads.
Every message is attributed
Agent name, the model it used and the sources it drew on. Nobody has to guess where an answer came from.
Anything can be pinned
Pin an output and it is indexed into the Space's knowledge, so the next question that touches it gets a grounded answer.
AI Agents
Agents that belong to the team, not to whoever is talking to them.
An agent added to a Space has a name, a job and a permission set. It acts only where the Space's grants and its own connector permissions overlap, and it can be active in one Space and paused in another.
A named manager, accountable
Every agent has a manager who is its only editor and is accountable for re-auth, rejected proposals and spend. Other members request a change. That manager's ID is on every audit line the agent produces, and being the manager grants no access to what the agent read.
Scoped per connector, not blanket
Read or write per agent per connector. Revoke in seconds.
It proposes, it does not publish
An agent either proposes changes to a knowledge base for approval, or reads it only.
Added from inside the Space
An Owner adds and edits agents without leaving the team's workspace.
Workflows
The repeatable work, running on its own.
Build a multi-step automation on a node canvas. Run it on a schedule, on an external event, or manually. A step can run an agent, in which case that agent's own permissions and manager still apply.
Visibility
Open by default, restricted in one click.
Open Spaces let anyone in the organization find and join, which is how good work spreads past the team that started it. One click makes a Space invite only, and an admin can make restricted the default organization-wide.
Attaching knowledge to an open Space shows the owner exactly who can see it before they attach it. Roles stay simple: Owner and Member inside a Space, under organization roles of Admin and Member. Whoever creates a Space is its first Owner.
Organization admins can see every shared Space, restricted ones included, so nothing ungoverned grows in a corner.
Surfaces
The people who never open a new app still take part.
Link a Space to a Slack channel. Mention OrgLogic in the channel and the thread opens or continues in the linked Space, and Space threads mirror back.
Only organization members mapped to a Slack identity can read or post through the link. Also on the web app, Microsoft Teams and Chrome, on every plan including Free.
And for the people who would rather stay in Claude or ChatGPT: connect the Company Brain over MCP and they retrieve what the Spaces they belong to know, inside the client they already use.
The boundary
What is in a Space is known by the Space. Nothing else is.
A thread is grounded by its Space's knowledge and memory. Nothing appears in a Space unless the Space was granted it or a member deliberately put it there. That single rule is what lets a company hand Spaces to every team without a review for each one.
Personal stays personal
A person's own connections, the knowledge created through them and their personal memory stay at user level in their Private Space. Sharing into a Space is a deliberate act, attributed and logged.
Knowledge follows its source
If the member who synced a source leaves the Space, that knowledge disconnects and the Owner is notified. Any member reconnecting the same source restores it. Native content stays with the Space.
Every message on the record
Thread messages and agent actions logged with Space, thread, actor, agent, its manager, model, sources and cost. Searchable and exportable.
Redacted before the model
PII detected and masked before any prompt reaches a model, from every Space and every agent.
Cost visible per team
Budget controls with alerts and hard caps. Spend by team, user, model and agent.
Nothing trained on your data
Contractual, whichever way you route. SOC 2 Type II, ISO 27001, GDPR, HIPAA BAA.
In practice
Teams get to daily use in about two weeks
A pre-opening cognitive city development in Florida, fourteen districts, built by a confidential enterprise, reached a 2-week median team adoption, department to daily active use, via Slack and Teams. It now runs 14 agents in production across finance, legal, procurement, investor relations and program management, with CISO sign-off in month one.
Two weeks matters more than it looks. Single-player AI tools roll out for months and still serve the same handful of power users, because every person has to discover their own value from scratch. When the value is already in the Space, the next member inherits it.
2 wks
Median team adoption
Department to daily active use, via Slack and Teams.
14
Agents in production
Across finance, legal, procurement, investor relations and program management.
Month 1
CISO sign-off
Governance shipped as the first deliverable.
Pricing
Unlimited Spaces on every plan.
Free
$0
Up to 25 users, no seat minimum, no card.
Unlimited Spaces, threads and agents
Full governance
$500 in pooled model credits
Standard
$8
seat / mo
Up to 25 users, no seat minimum, no card.
Everything in Free
API and SDK
Enterprise
Custom
100 seat minimum.
SSO, SCIM, domain verification
Single-tenant VPC and on-premise
FAQ
Spaces FAQ
What is a Space?
A container for a team's work with AI. It owns the team's members, both people and agents, its threads, its knowledge bases and its memory.
How many Spaces can we create?
Unlimited, on every plan including Free. Any member of the organization can create one.
Who can see a Space?
Open Spaces can be found and joined by anyone in the organization. Restricted Spaces are invite only, and an admin sets which is the default. Organization admins can see every shared Space, restricted ones included.
What is the difference between a Space and my Private Space?
Your Private Space is the same container with one human in it: your own threads, your own connections, your own knowledge and memory. Sharing something out of it into a Space is a deliberate act, and it is attributed and logged.
Does my personal knowledge end up in a Space?
Only if you put it there. A person's own connections and the knowledge created through them stay at user level until they deliberately share.
How do agents get permissions in a Space?
An agent acts where the Space's grants and its own connector permissions overlap. Read or write scope is set per agent per connector and revocable in seconds. Each agent has a named manager who is its only editor.
What happens when someone leaves the Space?
Knowledge synced through their connection disconnects and the Owner is notified, and any member reconnecting the same source restores it. Thread files, pinned outputs and everything native to the Space stay.
Can we work in Slack instead?
Yes. A Space links to one Slack channel, mentions in the channel continue the thread in the Space, and Space threads mirror back.
Can a team start without IT?
Yes. Sign up free, create a Space, invite a teammate. The Assistant is already a member and governance is already on.