Anthropic expands Claude Marketplace with enterprise tools

Anthropic has added more than 25 partner products to Claude Marketplace, with purchases possible through clients' committed enterprise budgets. At the same time, Claude can now verify user identity for business plugins.

On November 4, 2026, Anthropic expanded Claude Marketplace, its enterprise marketplace, with over 25 partner agents and products covering security, developer infrastructure, creation, and productivity. At the same time, the company introduced a new authentication feature that allows Claude to verify a user's membership in an organization and automatically connect certain business plugins. The goal goes beyond user convenience: Anthropic aims to make Claude both an entry point for software purchases and for access control. Not just a chatbot, then.

A marketplace aiming to influence software purchasing

Claude Marketplace does more than just add applications to an existing catalog. Anthropic now distinguishes three blocks: connectors and plugins that bring tools into Claude, service partners who assist with deployment, and partner products that can be purchased separately. This separation matters because it brings Claude closer to a more traditional software purchasing process, with references, use cases, and a procurement framework better defined for IT departments.

The model mainly targets enterprise clients. Anthropic states that eligible buyers can use part of their Anthropic spending commitment to finance these partner products, including new purchases or extensions on platforms already in place. In other words, a single budget can cover Claude and associated third-party software. For purchasing departments, this can simplify accounting. For partner publishers, it opens a door to a distribution channel already connected to AI budgets.

Which products are included in the catalog?

The catalog highlights tools focused on security and infrastructure, as well as creation and productivity services. Chainguard, for example, builds applications from source code in an isolated microVM and offers remediation recommendations; Checkmarx One combines rule-based analysis and AI reasoning to detect vulnerabilities before production deployment. Cyera Agent Guardian governs Claude deployment permissions, Rubrik Code Guardian inspects GitHub or Azure DevOps repositories, and Tanium Cloud reports vulnerabilities, configuration errors, and missing patches in real time.

On the engineering side, Datadog Agent Observability traces agent execution across tools and services, while Netlify Enterprise handles deployment on a global edge network with preview URLs, SSO, roles, and audit logs. The creation and productivity block includes AirOps, Runway, Webflow, Asana, Glean, monday, and Zoom, with uses ranging from video generation to enterprise information search. Overall, the logic is simple: Claude no longer just wants to respond, it also wants to be the counter where you choose the building blocks around it.

Claude also becomes an identity gatekeeper

The second announcement concerns digital identity. Anthropic now says it can use Claude itself to issue signed identity assertions, which compatible services can accept to recognize an employee. This approach complements the Enterprise-Managed Authorization (EMA) extension of the MCP protocol, introduced in June, which already allowed centralized management via a compatible identity provider. The problem was that EMA adoption remained limited outside Okta, the first provider to support it.

With this new option, Anthropic bypasses part of the blockage. Team and Enterprise administrators can choose authorized services and relevant roles. When a user opens an eligible plugin, Claude sends a signed assertion proving they belong to the organization; the service then verifies this assertion before issuing an access token. The connection is then automatically renewed when tokens expire. Anthropic notably cites more than a dozen services at launch, including Amplitude, Atlassian, Datadog, Dropbox, Figma, HubSpot, Linear, Notion, PagerDuty, Semrush, and Supabase.

What this means for companies

The sensitive point is quite clear: Anthropic is approaching a function traditionally reserved for identity providers, without having all the responsibilities or guarantees. Connected services keep their own access rules, and Supabase reminds that a corresponding SSO account must already exist on the company side. Above all, revocation is not just about removing a user from the internal directory: the administrator must also delete them on Claude and on the third-party service. The kind of detail that can prevent headaches... or create another if no one follows the process.

The scope remains limited to the enterprise perimeter and beta. But the signal is clear: Anthropic seeks to position itself at the crossroads of access, provisioning, and software purchasing. For clients, this can reduce friction. For security and IT teams, it adds a layer to manage. And for partner publishers, it can become an interesting commercial channel, provided they align with Claude's technical rules and enterprise account requirements.

Key points

  • More than 25 partner products in Claude Marketplace.
  • Launch announced on November 4, 2026 by Anthropic.
  • Claude can sign user identity for certain plugins.
  • Okta remains the first provider to have supported EMA.
  • Supabase cites Claude as a possible new issuer.
  • Beta reserved for Team and Enterprise clients.

By the numbers

  • 25+ products — Claude Marketplace partner catalog, November 2026, Anthropic.
  • 14 services — launch of auto-connect plugins, November 2026, Anthropic.
  • 1 provider — Okta, first EMA support cited by Anthropic in June 2026.

To read