Anthropic ships Claude Managed Agents, raises Claude Code API limits at Code w/ Claude
Multiagent orchestration, dreaming, webhooks. Self-hosted sandboxes in public beta, MCP tunnels in research preview.
Anthropic opened Code w/ Claude in San Francisco on Wednesday with the public release of Claude Managed Agents — a hosted platform that bundles multiagent orchestration, scheduled “dreaming” runs, Outcomes telemetry, and webhooks on top of the Claude API.
Claude Code and Opus API rate limits went up the same day. Self-hosted sandboxes entered public beta for teams that need to keep sensitive files, packages, and services in their own infrastructure. MCP tunnels — for exposing tools inside a private network to an external Claude client without inbound ports — entered research preview.
On the vertical side: more than 20 new MCP connectors for legal-research systems and 12 practice-area plugins covering research, contracts, discovery, matter management, and legal aid.
Anthropic also expanded Project Glasswing with Claude Security in public beta and new cyber verification tools for eligible security teams.
The conference runs in London on May 20–21 and Tokyo on June 5–6. Simon Willison is keeping a live blog of the announcements, which is a useful sequence-of-events reference for readers who can’t watch the keynotes in real time.
Read on the surface, the message of the day is that the scaffolding-as-product layer has consolidated. Six months ago “build a production agent on Claude” meant assembling half a dozen self-managed services around the API call. As of Wednesday, most of that scaffolding sits inside Anthropic’s plane.