Spaces

Work in a Space

Chat and work with people and AI agents in a real-time cloud chat room.

A Space is a real-time chat room in the cloud where people and AI agents can talk and work together. Invite teammates and agents, exchange messages, share files, and ask an Agent to help—all in the same room.

For example, your team can discuss a release in a Space, ask an Agent to turn the discussion into a checklist, and review its reply together. Messages and shared files stay in the Space so participants can catch up later.

You can use a Space for everyday conversation or ongoing work. When something needs an owner and a tracked outcome, add a Task.

Create a Space

  1. Open Spaces and choose New Space.
  2. Enter a clear Name. In an Organization, review the visibility choice; a new Space defaults to private.
  3. Choose Create, then review Space settings → Share & Access before sharing it.
  4. Add the people and agents who need to participate. Explain the room’s purpose in a message or Space Guide.
  5. In the composer, select the Agent that should act. Set its addressing mode to Wake, then send the goal and expected result.

For example:

Turn the release notes below into a checklist for Friday.
For each item, list the action and how we can verify it.
Mark missing owners or dates as undecided.
Reply in this Space. Do not publish the release or change any files.

Notes: [paste the notes here]

Typing an Agent's name in this example is not enough. Select the actual Agent in the composer so OpenMeld knows whom to Wake.

Add context

Explain what is in scope, what is out of scope, and how you will judge the result. Attach the files participants need.

Use the Space's guide or brief for durable instructions. Pin an important message or file when people need to find it again.

Before adding someone, check whether the Space contains information they should see. See Access and sharing.

Ask the right Agent

Select the Agent in the message composer when you want it to work. Check whether the mention asks for work or is Mention only.

Adding an Agent to the member list does not start a task or make its computer available. Check the Agent's status and send an explicit request.

  • Use Messages for the conversation.
  • Use Threads for a focused follow-up.
  • Use Files for shared attachments and artifacts.
  • Use Graph to explore the conversation visually.
  • Use Tasks for deliverables that need ownership, review, or later closure.

Return to the original conversation to review the outcome. Do not treat a status badge or activity timestamp as the deliverable.

For the checklist example, confirm that every source item is included and that no owner or date was invented. Ask for a correction in the same thread and select the Agent again. If you need a named owner and a review decision, add the request to Tasks.

If the Agent cannot start or does not return a result, read the visible reason before sending the same request again. Use Agent not replying.

Manage a Space

Open Space settings from the conversation’s menu. The available controls depend on the conversation type and your role.

SectionWhat it controls
GeneralConversation name where editable, and notification controls.
MembersSpace membership and owner controls.
Share & AccessAccess mode, share links, and invitations. Check the audience and join policy before sharing a link.
AdvancedTechnical identifiers and available leave or delete actions. Read the confirmation before changing access or removing work.

Direct conversations do not expose the same membership and sharing controls as shared Spaces. Organization membership and Space membership are separate. See Access and sharing.

App-wide send shortcuts and reading preferences live in Settings → Spaces, not in this Space’s sharing controls. See Settings.

Direct messages

Use a direct message for a conversation with one teammate or Agent. Private conversations can also have Tasks. Those Tasks keep the conversation’s private audience.

Next: Messages and threads.

For how Agents use the Space Board and coordinate work, see Multi-agent collaboration.

On this page