Today ainotis Join

My notis

ToolingPublished All news from that day

Anthropic puts self-hosted Claude Managed Agents in private beta, so companies can run it on their own infrastructure.

In private beta from 12 November, customers run the server, database and workers and pick where model calls go, such as Bedrock, Google's Agent Platform or Microsoft Foundry.

By ai notis. An AI newsroom with a human editor. Every fact links to its source. How we work

Share

Check our sources · 13 facts from 4 sources

Anthropic put self-hosted Claude Managed Agents into private beta on 12 November, so a company can run the whole Managed Agents service, not only the sandboxes, in its own cloud account or on-premise. Anthropic announced it in a post on 12 November and documents it on the Claude Platform docs site.

The customer runs the Managed Agents server, a PostgreSQL 16 or later database, and the workers that execute the agent's shell and file tools, and session history is stored in that database. Model requests go where the customer chooses: Anthropic directly, Amazon Bedrock, Google's Agent Platform, Microsoft Foundry or its own gateway. Access is by request through an Anthropic account team.

The docs say the server has the same API as the hosted service, with listed exceptions. Cloud environments, the Messages and Models APIs, the Files API, webhooks, scheduled deployments and the prebuilt docx, pdf, pptx and xlsx skills are not available on the self-hosted server.

The security work moves to the customer. The docs say the agent's bash tool is a real shell with the worker's permissions and network access, that isolation comes from what the worker runs in, and that what Anthropic's sandbox page assigns to Anthropic, such as securing the server and detecting unusual usage, is the customer's responsibility.

On data retention, Anthropic's post says an organization's zero data retention arrangement applies to requests sent to Claude in the same way as any other Messages API call. Anthropic's retention page lists Managed Agents as not eligible for ZDR, with self-hosted Managed Agents as the exception because the customer stores the transcripts.

Share this story

Your reaction

We count reactions per story and day, never who reacted. The counts help us choose what goes in the monthly issue. If you are signed in, your own page shows yours too.

Check our sources

Every sentence above is checked against these 4 sources.

1 Operate Claude Managed Agents inside your security perimeterAnthropic · 12 Nov 2026 · 4 facts Open the source archived copy

  1. Anthropic's post of November 12, 2026 says self-hosted Claude Managed Agents is now available in private beta.

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

  2. Anthropic's post says model requests go where the customer chooses: Anthropic directly, a cloud provider (Amazon Bedrock, Google's Agent Platform, or Microsoft Foundry), or the customer's own gateway.

    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.
    cite

  3. Anthropic's post says that if an 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.

    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.
    cite

  4. Anthropic's post says that in self-hosted Managed Agents session history is stored in the customer's own database.

    Session history is stored in your own database.
    cite
2 Self-hosted Claude Managed Agents (docs)Anthropic · 6 facts Open the source archived copy

  1. Anthropic's docs say self-hosted Claude Managed Agents is Managed Agents on the customer's own infrastructure, which can be a cloud account or an on-premise data center, and that the customer hosts the service and the sandboxes.

    Self-hosted Claude Managed Agents is Managed Agents on your own infrastructure, which can be a cloud account or an on-premise data center. You host the service and the sandboxes.
    cite

  2. The docs say the self-hosted server (cma-server) serves the API, runs the agent loop, and stores agents, sessions, and each session's history, and that workers take work from the server and run the agent's shell and file tools in the customer's sandboxes.

    Serves the API, runs the agent loop, and stores agents, sessions, and each session’s history. ... Programs that take work from the server and run the agent’s shell and file tools in your sandboxes.
    cite

  3. The docs say the database is PostgreSQL 16 or later, and that without a database the server keeps its data in memory and a restart erases it.

    PostgreSQL 16 or later. Without a database, the server keeps its data in memory, and a restart erases it.
    cite

  4. The docs say that to request access to self-hosted Claude Managed Agents, customers contact their Anthropic account team.

    To request access to self-hosted Claude Managed Agents, contact your Anthropic account team.
    cite

  5. The docs say the agent's bash tool is a real shell with the worker process's permissions and network access, that the worker does not isolate the shell, and that isolation comes from what the worker runs in.

    The agent’s bash tool is a real shell with the worker process’s permissions and network access. The worker doesn’t isolate that shell. Isolation comes from what the worker runs in.
    cite

  6. The docs say that what Anthropic's self-hosted sandboxes security page assigns to Anthropic, including securing the server, hosting memory stores, and detecting unusual usage, is the customer's responsibility in self-hosted Managed Agents.

    What that page assigns to Anthropic is your responsibility, including securing the server, hosting memory stores, and detecting unusual usage.
    cite
3 Differences from the hosted service (self-hosted Managed Agents docs)Anthropic · 2 facts Open the source archived copy

  1. Anthropic's docs say a self-hosted server has the same API as the hosted service apart from listed differences, and that nothing is copied from the hosted service, session history included.

    A self-hosted Claude Managed Agents server has the same API as Anthropic’s hosted Managed Agents service, apart from the differences on this page. ... Nothing is copied from the hosted service, session history included.
    cite

  2. The docs list as not available on a self-hosted server: cloud environments, the rest of the Claude API such as the Messages and Models APIs, file session resources and the Files API, mcp_oauth vault credentials, the prebuilt docx, pdf, pptx and xlsx skills, scheduled deployments and dreams, and webhooks.

    Cloud environments ... The rest of the Claude API, such as the Messages API and the Models API ... File session resources and the Files API ... mcp_oauth vault credentials ... Prebuilt docx, pdf, pptx, and xlsx skills ... Scheduled deployments and dreams ... Webhooks
    cite
4 API and data retention (Claude Platform docs)Anthropic · 1 fact Open the source archived copy

  1. Anthropic's data retention page says Claude Managed Agents is not covered by ZDR because session transcripts persist until deleted, with the exception of self-hosted Claude Managed Agents, where the customer stores the transcripts.

    cite

Topics

The morning email

On the mornings we publish: the three top stories and up to four short ones. Free.

We email you a link to confirm. An issue may include one sponsor, always labelled Sponsored · Advertisement. Our emails count opens and clicks, not who made them. Unsubscribe in one click. What we keep