Claude Platform

Developers and operators can use the new server-evaluation mode to automate or gate agent tool usage and use the CLI/session viewer to monitor and approve tool calls in real time.

Availability

Beta for Claude Enterprise organizations; not available to Claude Platform (Claude Console) organizations

Pricing model

Newly billed refusals are charged like any other request, at the rates of the model that ran it

Deployment

Hosted via Claude API

Privacy

server evaluates tool calls and can pause for human approval

Licence

Requires an Admin API key with the read:claude_tag or write:claude_tag scope

• Managed Agents permission policies now include an "auto" mode that lets the server evaluate each agent or MCP tool call and choose to run it, deny it, or pause for approval. ^[1]

  Claude Managed Agents permission policies now include auto : the server evaluates each agent or MCP tool call and runs it, denies it, or pauses for your approval.

• agent.tool_use and agent.mcp_tool_use events now report an evaluation field alongside evaluated_permission describing how each call was evaluated. ^[1]

  agent.tool_use and agent.mcp_tool_use events report how each call was evaluated in an evaluation field alongside evaluated_permission .

• The ant CLI v1.32.0 adds ant beta:sessions connect to attach a terminal to a Claude Managed Agents session for live following and interaction. ^[1]

  Version 1.32.0 of the ant CLI adds ant beta:sessions connect , which attaches your terminal to a Claude Managed Agents session.

• The CLI command supports sending messages and allowing or denying tool calls that are waiting for approval. ^[1]

  You can follow the session live, send messages, and allow or deny tool calls that are waiting for approval.

• Passing --web to the CLI serves the Claude Console session viewer locally and opens the session there. ^[1]

  Pass --web to serve the Claude Console's session viewer locally and open the session there instead.

• The Messages API can compact a conversation on demand in beta using the compact-2026-09-04 header. ^[2]

  The Messages API can now compact a conversation on demand on the Claude API, in beta with the compact-2026-09-04 beta header.

• The API returns a signed compaction block that summarizes the messages you sent, which can be sent in later requests in place of those messages. ^[2]

  Send the top-level compaction parameter, and the API returns a signed compaction block that summarizes the messages you sent.

• Compaction can run in the background and allows keeping recent turns word for word after the summary. ^[2]

  You choose when to compact, the request can run in the background, and you can keep recent turns word for word after the summary.

• On models with preserved thinking, the thinking in kept turns can remain valid after compaction. ^[2]

  On models with preserved thinking, the thinking in those kept turns can stay valid.

• With the thinking-binding-controls-2026-08-01 header, input_transformations gains thinking_mismatch_allowed entries that name thinking blocks that failed prefix checks when enforcement is not enabled. ^[2]

  With the thinking-binding-controls-2026-08-01 beta header, the input_transformations response field gains a second entry type, thinking_mismatch_allowed .

• The thinking_mismatch_allowed block still reaches the model unchanged and can be logged to find history edits before opting into enforcement. ^[2]
• Pre-output refusals with stop_details.category "bio", "frontier_llm", or "reasoning_extraction" are now billed. ^[3]

  We're expanding which refusals are billed to include refusals that arrive before any output when stop_details.category is "bio" , "frontier_llm" , or "reasoning_extraction" , the categories where we measure low volumes of false positives.

• Newly billed refusals are charged at the rates of the model that ran the request. ^[3]

  The newly billed refusals are charged like any other request, at the rates of the model that ran it.

• Pre-output refusals in other categories are still not billed. ^[3]

  Refusals before any output in other categories are still not billed, and fallback credit is unchanged.

• Fallback credit is unchanged. ^[3]

  Refusals before any output in other categories are still not billed, and fallback credit is unchanged.

• The change applies on all platforms. ^[3]

  This change applies on all platforms.

• The Claude SDK emits spans in Python, TypeScript, C#, Go, Java, PHP, and Ruby. ^[4]

  The Claude SDK emits spans in Python, TypeScript, C#, Go, Java, PHP, and Ruby.

• OpenTelemetry export supports tracing Messages API requests and Claude Managed Agents sessions. ^[4]

  We've launched OpenTelemetry export in beta, so you can trace Messages API requests and Claude Managed Agents sessions in your observability backend.

• OpenTelemetry export is available in beta. ^[4]

  We've launched OpenTelemetry export in beta, so you can trace Messages API requests and Claude Managed Agents sessions in your observability backend.

• To get spans from the API, users can add an OpenTelemetry destination in the Claude Console. ^[4]

  To get spans from the API, add an OpenTelemetry destination in the Claude Console.

• The endpoint is GET /v1/compliance/organizations/{org_uuid}/user_preferences. ^[5]

  Call GET /v1/compliance/organizations/{org_uuid}/user_preferences with your existing Compliance Access Key and the read:compliance_user_data scope.

• The endpoint returns text saved as Instructions for Claude in claude.ai and as Global instructions in Cowork. ^[5]

  The Compliance API now returns the text each organization member saved as Instructions for Claude in claude.ai and as Global instructions in Cowork, in beta for Claude Enterprise organizations.

• The endpoint is in beta for Claude Enterprise organizations. ^[5]

  The Compliance API now returns the text each organization member saved as Instructions for Claude in claude.ai and as Global instructions in Cowork, in beta for Claude Enterprise organizations.

• Calls require an existing Compliance Access Key and the read:compliance_user_data scope. ^[5]

  Call GET /v1/compliance/organizations/{org_uuid}/user_preferences with your existing Compliance Access Key and the read:compliance_user_data scope.

• Inference hooks are available for Claude Enterprise organizations. ^[6]

  Inference hooks are out of beta for Claude Enterprise organizations.

• Inference hooks are out of beta. ^[6]

  Inference hooks are out of beta for Claude Enterprise organizations.

• The release note concerns inference hooks on the Claude Platform. ^[6]
• Anthropic says the new SDK and CLI versions send API version 2026-10-21 by default. ^[7]

  We've released new versions of the client SDKs and the ant CLI that send API version 2026-10-21 by default: Python SDK 2.0.0, TypeScript SDK 1.0.0, C# SDK 13.0.0, Go SDK 2.0.0, Java SDK 3.0.0, PHP SDK 1.0.0, Ruby SDK 2.0.0, and ant CLI 2.0.0.

• User profiles include the enforcement_status and enforcement_actions fields. ^[7]

  We've added the enforcement_status and enforcement_actions fields to user profiles , so you can see whether Anthropic has restricted or banned one of your end customers.

• Anthropic released API version 2026-10-21. ^[7]

  We've released API version 2026-10-21 .

• The enforcement_status[] filter finds profiles with an open restriction or ban. ^[7]

  To find the profiles that have an open restriction or ban, list your profiles with the new enforcement_status[] filter.

• The API is not available to Claude Platform (Claude Console) organizations. ^[8]

  It is not available to Claude Platform (Claude Console) organizations.

• Calls require an Admin API key created by the primary owner in claude.ai with the read:claude_tag or write:claude_tag scope.
• The Claude Tag API is available in beta for Claude Enterprise organizations. ^[8]

  The Claude Tag API is available in beta for Claude Enterprise organizations with the experimental-claude-tag-2026-10-21 beta header.

• Its endpoints read and change Claude Tag settings on Slack channels and workspaces.
• The API shows which settings apply in a channel and where each comes from, and can attach Access bundles.

• Create expressive speech from a short voice sample Daily brief 2026-10-27
• Access Claude through a Google Cloud project Daily brief 2026-10-24
• Extract structured text from PDFs and images on a CPU Daily brief 2026-10-14
• Claude Platform adds automatic evaluation for Managed Agents tool calls Daily brief 2026-09-24