The Black Cloud

AI - Claude Meetings joins your calls, and Claude Managed Agents can now be self-hosted

I said it was not my usual thing to do Anthropic news two days in a row… and here we are with the third Anthropic post this week 🤣

When I wrote about Claude Whiteboards, Sheets and Data, Claude Meetings was just a “coming soon” line at the bottom of the announcement. Well, “soon” turned out to be two days 😊 Claude can now join your video call, take the notes and build docs and whiteboards with you while you are still talking. On the developer side, Anthropic has also opened a private beta that lets you run the whole Claude Managed Agents service in your own cloud or on-premises.

And yes, this landed the day after I wrote about Grok Bot 1.0 joining meetings, so the AI-in-your-meeting race is very much on 😉

Claude Meetings

Claude Meetings adds Claude to a Google Meet call as a visible participant. It is not just another notetaker, though. The idea is that the work starts during the meeting: you ask Claude to start a doc or a whiteboard, and your team can edit it while the conversation carries on.

Official announcement

Introducing Claude Meetings: add Claude to your team’s video call and it builds docs, whiteboards, and other artifacts with you while you’re still talking.

Starting to roll out in beta on Team and Enterprise plans.

Introducing Claude Meetings, now in beta

Claude can now join your video calls as a visible participant, follow the conversation and start on work during the meeting. Claude Meetings is now in beta.

Your team can ask Claude questions during the meeting, out loud or in chat. By default, Claude answers in writing, and it can use your connected tools with your permission.

Claude also works with you live to build docs, whiteboards, or other artifacts, while teammates on the invite refine them. Claude writes the notes and action items, so no one on the call has to.

Claude Meetings is starting to roll out in beta on Team and Enterprise plans. Support for other video conferencing platforms is coming soon.

What is new?

According to Anthropic’s announcement and the Claude Help Center:

NOTE

Claude Meetings is rolling out in beta on Team and Enterprise plans in Claude and Claude Desktop, and it works with Google Meet only for now (Anthropic says other platforms are “coming soon”). While it is in beta, it is not available to organisations with zero data retention (ZDR), the HIPAA-ready configuration or customer-managed encryption keys (CMEK).

Defaults for admins: on Team plans, Claude Meetings is on by default; on Enterprise plans, it is off until an owner turns it on for all or part of the organisation (using custom roles). Owners manage it in Organization settings > Products > Claude Meetings (beta), which also controls calls with outside guests and whether Claude can use members’ connectors in meetings.

Outside guests: Anthropic’s announcement says Claude stays out of calls with outside guests by default, but the Help Center says that on Team plans Claude can join calls with outside guests by default, and only Enterprise requires an owner to allow it. If that matters to you, check the setting rather than relying on the announcement. When guests are allowed, what they say is included in the notes and transcript, although they do not get the meeting page or the recap email.

Pricing during the beta: notes, transcripts and answers that do not use your connected tools are included on Team and Enterprise plans. Docs and whiteboards made in a meeting are free until 10 February 2027. Anything that uses your connected tools counts towards your usage and can lead to charges for your organisation.

To try it: go to Meetings in Claude, connect Google Calendar (an owner has to enable Google Calendar for the organisation first), select a meeting and click Add Claude. Auto-join is in Settings > Capabilities > Meetings > Auto-join. And please make sure everyone on the call is OK with being transcribed before you add it.

From the people who built it

Claude can now join your meetings as an active participant.

It’s unlike anything I’ve used on a call, it can:

• Build a doc or whiteboard live as you talk
• Check connected tools with your OK
• Answer the team when asked out loud

The work kicks off before the call ends.

claude meetings is that girl!!!! ask claude to make to make a live whiteboard during your call and watch it map out what you’re talking about in real-time 😮

A whiteboard that draws itself while people argue about the timeline is genuinely the bit I would like to see in action 😊

What people are saying

The best summary of the week

Corey Ganim put the whole week of Anthropic launches into context, and it is worth reading in full:

Claude shipped Whiteboard, Sheets, Data, and Meetings in 3 days. My raw thoughts:

  1. This is the Microsoft Office period for AI. Office bundled Word, Excel, and PowerPoint, basically killing Lotus 1-2-3 and WordPerfect.

  2. Notetakers are going to feel the most pressure. Notes, action items, and transcripts are free on Team and Enterprise during the beta. Hard to charge per seat for that now.

  3. Something I find very interesting: Anthropic is “partnering” with Granola, Fireflies, Otter, and Gong. Claude records the meeting and they become where the transcript gets sent.

  4. Gong is safe (for now). Claude stays out of any call with an outside guest by default (i.e. sales calls)

  5. Doing real work inside a call is the product here. Being able to say “Claude, start a doc for this plan” that your team can edit mid-call is a real efficiency gain.

  6. Interactive connectors are one of the recent updates nobody is talking about. Adobe, Atlassian, and Box now open INSIDE Claude. It seems that Anthropic knows which tools it can’t replace. They’re choosing to host those instead.

  7. The pattern is clear, they’re coming for just about everything. Docs, boards, sheets, dashboards, meetings, and your bank balances (Personal Finance shipped Nov 5). Who knows what’s next?

  8. I think vertical SaaS is Anthropic’s next play. Claude for Legal got a contract review gallery last month while Claude for Manufacturing launched today with 17 skills for quotes and quality reports.

  9. Meetings ships inside Google Meet only and on Team and Enterprise only (for now, in beta). If you’re a HIPAA or zero-data-retention org, you shouldn’t touch it yet. Zoom shops and healthcare will have to wait.

  10. Perfect time for your org to audit your stack and figure out which tools you’re currently paying for that are now bundled in with your Claude subscription.

  11. Even with all these new features, Anthropic has done a great job in keeping the Claude app SIMPLE. This is where OpenAI has really dropped the ball. The ChatGPT app is incredibly complex and confusing by contrast.

Such an incredible time to be a builder and on the forefront of AI. Remember, most of humanity still has zero grasp of these tools!

I agree with most of this, especially points 5 and 11. Two small corrections, though. On point 4, that is only the default on Enterprise plans; according to the Help Center, on Team plans Claude can join calls with outside guests by default, so “Gong is safe” depends on how your admin has set things up 😉 And on point 9, it is not so much “should not touch it” as “cannot”: the beta is simply not available to ZDR, HIPAA-ready or CMEK organisations yet. I would also add “Microsoft Teams shops” to the list of people who have to wait.

Claude in your meetings, or your meetings in Claude?

Today → Claude in your meetings

Tomorrow → Your meetings in Claude

Short, but I think Merill has nailed where this is heading. Once the meeting page, the doc, the whiteboard and the follow-up chat all live in Claude, the video call itself becomes the least important part 🤣

The reactions

Claude meetings introduced. Joins calls and builds docs, whiteboards and other Artefacts.

Ngl, looks really useful. No GPT-6.2 today, but Claude saved the day (a bit at least)

Another day, another 100 startups gone 💀

My favourite reaction, though, came in Turkish:

Translated from Turkish:

Bad news for those friends who join the meeting with their camera off and do not say a single word for an hour.

Claude has come to replace you. Its camera is off too, and it also sits quietly in its corner. The only difference is that it actually listens to the meeting.

It takes notes, writes the document, and when it leaves, it does not ask “Did I miss anything?”

We have all been in that meeting 🤣

Self-hosted Claude Managed Agents

The second announcement is aimed at developers and platform teams rather than everyday Claude users. Claude Managed Agents is Anthropic’s set of APIs for building and running long-running autonomous agents: it provides the agent harness, sandboxes, credential handling and session history. Until now, Anthropic ran the service, and the most you could host yourself was the sandbox where an agent’s tools run. Now, in a private beta, you can host the whole thing.

Official announcement

Self-hosted Claude Managed Agents is now in private beta.

You can host the agent harness, sandboxes, and session history, in your own cloud or on-premises, and choose where requests to Claude are sent.

Operate Claude Managed Agents inside your security perimeter

With self-hosted Managed Agents, every part of the service runs in your own cloud or on-premise, including the harness, sandboxes, credentials, and session history. You operate the Managed Agents server, its database, and the workers where agents run their shell and file tools. Model requests go where you choose, whether that’s Anthropic directly, your cloud provider (Amazon Bedrock, Google’s Agent Platform, or Microsoft Foundry), or your own gateway. Your developers still build agents with the same core APIs as the hosted Managed Agents service.

Session history is stored in your own database. If your organization has zero data retention (ZDR) in place with Anthropic, it applies to the requests sent to Claude in the same way as any other Messages API call.

What is new?

According to Anthropic’s announcement and the Claude Platform documentation:

NOTE

Self-hosted Claude Managed Agents is in private beta. To request access to the download, contact your Anthropic account team. The server software comes under its own licence agreement, and I could not find any published pricing.

It is not identical to the hosted service. According to the differences page, there are no cloud environments, no Files API, no prebuilt docx, pdf, pptx and xlsx skills, no scheduled deployments and no webhooks, and the Claude Console does not show what is running on your server. You also run, secure and update the server, its database and your sandboxes yourself, and each install of a new release is up to you.

Data still leaves your network for model calls: the server sends the conversation so far, including tool inputs and results, to your chosen inference provider (the Claude API by default).

What people are saying

Translated from Arabic:

A powerful update from Claude.

Now you can run AI Agents on your company’s servers and control the runtime environment and the data.

A big step for data privacy in companies.

The feature is currently in Private Beta.

#AI #Claude

Abdullah sums up why this matters. For a lot of organisations, the question is not whether agents are useful, but where the code, the credentials and the session history end up.

My take

Claude Meetings is the one most people will notice. Taking notes is now table stakes; Otter, Fireflies, Copilot in Teams and, as of this week, Grok Bot all do some version of it. The interesting part is that Claude starts the actual work while people are still talking, and that the doc or whiteboard is something the whole team can edit there and then rather than a summary that lands in your inbox an hour later.

That said, it is Google Meet only for now, and that is a big “for now”. Grok Bot already joins Google Meet, Zoom and Microsoft Teams, so until Claude supports Teams, a huge chunk of the business world (very much including the Microsoft shops I usually write about) will be watching from the sidelines. I would also pay close attention to the outside-guest setting: the difference between the announcement and the Help Center is exactly the kind of thing you want to check before Claude ends up transcribing a customer call 😉

Self-hosted Managed Agents is the less flashy announcement, but for regulated organisations it may be the more important one. Hosting the harness, the credentials and the session history yourself removes a big objection to letting agents loose on internal systems. Just remember that “self-hosted” also means “self-managed”: you run it, secure it, patch it and live with a slightly smaller feature set than the hosted service.

I have not tried either of them yet, but I will definitely be keeping an eye out for when Claude Meetings learns to join Microsoft Teams calls 😊

<< Previous Post

|

Next Post >>

#Blog #AI #Claude #Anthropic #Claude Meetings #Claude Managed Agents #Google #Google Meet #Agents #Security