Anthropic launches self-hosted Claude Managed Agents in private beta, running agents inside your own perimeter
krishnan · x · 2026-11-16
Anthropic announced Self-hosted Claude Managed Agents (private beta): companies can run the entire agent service—harness, sandboxes, credential vault, and session history—inside their own cloud or on-prem data center.
- Managed Agents was previously hosted-only; as agents take on critical work (fixing builds, testing releases, triaging alerts), teams want the whole control plane behind their security perimeter.
- Three tiers: cloud sandboxes (default), self-hosted sandboxes, and fully self-hosted Managed Agents where the model still runs at Anthropic or a provider you choose.
- A thread by krishnan highlights the fine print: self-hosted ≠ fully local—every model call sends the full conversation including tool inputs/outputs to the configured inference provider. Nothing migrates from hosted (including session history), no webhooks or scheduled deployments, and a first tool call silently waits if no worker is polling, so operators must watch for stalled jobs and test mid-job worker failures.
Related event: Anthropic Opens Private Beta for Self-Hosted Claude Managed Agents(4 posts)→
More from coding & agent
- Gemini 4 Hydrogen launched: 2.5x faster than 3.8 Flash, but how much is batching? — krishnan · 2026-11-16
- Running 13 agents in parallel: handhold the critical bits, let the rest rip all day — willcb · 2026-11-16
- Agent sessions stall silently when no worker polls—ops must watch for it — krishnan · 2026-11-16
- Self-hosted Managed Agents migrates nothing—not even session history—and drops webhooks and cron — krishnan · 2026-11-16
- Self-hosting agents buys control but you own the server, database, sandboxes and every release — krishnan · 2026-11-16
- Researchers formally specify ops interfaces for CockroachDB, Kafka and RabbitMQ to make AI operators verifiable — tianyin_xu · 2026-11-16