Anthropic is now allowing businesses to automatically connect employees to authorized Claude plugins, eliminating the need for staff to sign in and approve each service separately. The feature, announced Tuesday on the ClaudeDevs X account, enables Claude to directly confirm users' organizational membership, making it simpler for companies to control access to workplace applications. Available in beta for Team and Enterprise customers, the capability means Claude can issue the signed identity assertions that participating services use to recognize employees.

The new approach addresses a key limitation of the Enterprise-Managed Authorization extension to the Model Context Protocol, which launched in June. While that extension allowed companies to manage connections centrally through existing identity providers, Okta was the only provider supporting it at launch and most still don't, according to a November 3 blog post from Supabase, one of the participating services. Organizations using other providers no longer have to wait because Claude can now serve as the issuer instead. The initial rollout includes more than a dozen services: Amplitude, Apollo.io, Atlassian, Datadog, Dropbox, Figma, Granola, HubSpot, Linear, Mixpanel, Notion, PagerDuty, Semrush, and Supabase. When an eligible user opens a plugin, Claude sends a signed assertion identifying them as an organization member, the receiving service validates that assertion before granting an access token, and the process handles future connections automatically without repeated authorization.

According to Boris Cherny, head of applications at Anthropic, "Signing in to your company's plugins one by one is annoying." Administrators can select which supported services to activate and which employee roles should get access—for instance, giving all staff access to a general productivity tool while limiting access to company databases to specific roles. Supabase explains in its blog post that customers can now use Claude to authenticate employees accessing its server, even if their identity provider lacks extension support. The company notes that employees must already have a matching Supabase single sign-on account, and their existing permissions determine what they can access.

The feature builds on Anthropic's June announcement of support for the Enterprise-Managed Authorization extension in Claude, Claude Code, and Cowork, along with implementation guidance for developers building compatible servers. The approach originally relied on the company's identity provider being compatible with the extension, meaning many organizations still had to authorize connections individually when their providers didn't support it. By taking on a role traditionally handled by enterprise identity providers—vouching for users when they access third-party services—Claude is effectively stepping into identity verification territory. Technically, Claude acts as a trusted issuer for participating services, verifying that users belong to a particular Claude organization, while connected services remain responsible for enforcing their own access permissions. Supabase cautions that removing someone from the company's identity provider alone won't revoke Claude-issued access, and administrators must remove departing employees from both Claude and Supabase.

The announcement has already drawn positive reactions from users frustrated with repeated authorization prompts. One X user wrote, "This is soo good, tired of reauthorizing in every other session," while another eager for Stripe to join remarked, "I spend more time reconnecting Stripe than every other plugin combined." For now, Claude's identity verification extends only as far as the services willing to accept its assertions about user identity. The move reflects growing tension in enterprise software between convenience and traditional security boundaries, where speed often wins but accountability structures haven't yet caught up.