Claude can now join video meetings

On November 12, 2026, Anthropic began rolling out the Claude Meetings feature in beta for Team and Enterprise plans. This time, Claude isn’t an assistant waiting in a side panel, but a visible participant joining the video call. During the meeting, it answers questions, turns what’s being discussed into notes and action items, and even creates editable documents and whiteboards as the discussion moves forward.

The first supported platform is Google Meet. Anthropic says support for other video conferencing services is coming soon as well.

An assistant that works inside the meeting

Participants can address Claude both by voice and through chat. By default, Claude responds in writing; with the relevant person’s permission, it can use connected tools. In Anthropic’s example, a team planning a customer event asks Claude to start a plan based on what has been discussed so far. When the budget comes up, Claude asks for approval to pull last year’s cost, writes the figure into the chat, and compares it with the current plan. While the team moves the schedule onto a whiteboard, it keeps editing the document, and before the meeting ends, they can also ask for a summary.

This approach moves note-taking and draft writing from after the meeting into the middle of it. The practical benefit is turning the discussion into output while the context is still fresh.

A permanent page for every meeting

A separate page is created on claude.ai for each meeting: notes, transcript, and everything Claude generates stays there. Team members on the invite list can view the page, edit Claude’s outputs, and share them within the organization. Users can return to meetings they joined later and ask Claude questions about that meeting, and they can add Claude to a new meeting through the calendar.

Anthropic says it is working on integrations with Gong, Attention, Granola, Fireflies.ai, and Otter. These partnerships are planned for a later stage and are not part of the current beta.

Privacy and permission rules are decisive

When Claude joins a meeting, it appears as a participant and leaves a notification in the chat. It does not record audio; it stores transcripts, summaries, and screen descriptions. Anthropic wants users to make sure everyone has agreed to the transcript before adding Claude or enabling automatic joining.

The default behavior changes when someone from outside the organization is invited: Claude does not enter the meeting, and if an external participant joins later, it leaves the meeting. If an admin allows it, Claude can join meetings with external guests; in that case, it leaves a notification but does not respond in the chat, while the guest’s remarks still go into the notes and transcript. Connected tools work with each person’s own permissions, approval is requested on first use, and it remains valid afterward. For voice requests, Claude currently asks for confirmation every time to verify who is speaking. Everyone in the meeting can see the responses written in the chat.

Beta limits and pricing

Claude Meetings is currently unavailable for some organizations: accounts using zero data retention, a HIPAA-compliant configuration, or customer-managed encryption keys are outside the beta scope. On Team plans, meeting invites are enabled by default and admins can turn them off; on the Enterprise side, an admin needs to enable the feature organization-wide or partially.

During the beta, notes, transcripts, and in-meeting responses are free on Team and Enterprise plans. If a response uses a connected tool, it counts against the plan quota. Documents, whiteboards, and similar outputs created during the meeting will remain free until February 10, 2027.

Claude Meetings brings part of the work into the meeting itself, and moves part of the meeting into a managed workspace. That’s why the real value will depend not just on Claude’s document-writing ability, but on how transcript consent, guest policy, and tool permissions are set up.

Source: Unite.AI