Claude Code and Claude Cowork finally come to government!

Hi everyone, it's Shii! Today I found some news from a slightly more official corner of the world: Claude is now available for government agencies!

Claude Blog claude.com

What was announced?

On the Claude Blog, Anthropic announced that Claude Code and Claude Cowork are now available in public beta within Claude for Government Desktop. It's built on the same application commercial customers use, delivered through a FedRAMP High authorized environment. Claude Code helps public sector teams build and modernize software systems, while Claude Cowork works directly with files on the desktop, letting agency staff delegate memo creation, RFP reviews, casework, and presentations to Claude.

The story so far

Claude for Government already existed as a framework, but this update brings Claude Code and Claude Cowork into it, along with a broader set of government-specific governance capabilities. Spend tied to appropriated funds, department-level permission splits, and tamper-evident audit logs aren't things you typically find in a standard commercial plan, and government agencies need exactly this kind of infrastructure in place before they can adopt a tool like this.

What changes

This update brings three big changes:

• Feature velocity: agencies now receive new capabilities on the same cadence as commercial users. Conversation history is stored locally on agency-managed devices, and inference itself runs inside a FedRAMP High authorized environment.
• Billing that fits appropriations: program offices can use standard seat tiers or define custom ones with their own spend and model limits. Usage is purchased in fixed increments with a hard not-to-exceed cap, and the admin console shows per-user, per-model usage with automatic burndown alerts.
• Administration that matches departmental structure: department-level administrators can allocate seats and prepaid usage to sub-agencies, each of which manages its own users. SCIM group mappings set rate limits, spend caps, and allowed models per seat tier.

Dive Deep

Security and oversight get serious attention here too. Every administrative action is recorded in a hash-chained audit log that's reviewable directly in the product. Sensitive operations require two-person approval, and usage exports contain only metering data, so agencies can respond to ATO and Inspector General requests without moving sensitive material around.

Anthropic is also publishing its FedRAMP Secure Configuration Guide as a public document, so anyone can reference secure configuration practices. Formal change notifications required under FedRAMP, along with a penetration-test summary for the new desktop client, are available under NDA through Anthropic's trust center. The application itself deploys through agencies' standard MDM platforms.

Wrap-up

• Claude Code and Claude Cowork are now in public beta within Claude for Government Desktop
• Runs inside a FedRAMP High authorized environment, with conversation history stored locally on-device
• Adds government-specific governance: appropriations-based billing, departmental administration, and audit logging
• Sensitive operations require two-person approval; audit logs are hash-chained against tampering
• The FedRAMP Secure Configuration Guide is now public, and pentest summaries are available under NDA

If you care about the governance and procurement side of bringing AI into government work, this update is worth a look!