AI agents & Connected Apps - Stytch Docs

Agent-ready authentication

As AI agents gain the ability to access applications, Stytch Connected Apps provides the infrastructure to securely authenticate and authorize agent interactions—without rebuilding your auth system.

Core capabilities

Build an authorization server \ \Use Stytch’s pre-built UI components or headless SDK to create an OAuth-compliant authorization endpoint.

MCP server integration \ \Connect AI agents to your app via the Model Context Protocol with dynamic registration and scoped access.

Custom OAuth scopes \ \Define granular permissions based on real actions and resources in your app.

Client types \ \Configure first-party, third-party, confidential, and public clients with appropriate security controls.

Token management \ \Issue access tokens, refresh tokens, and ID tokens with built-in validation and revocation.

Consent management \ \View connected apps, authorized scopes, and revoke access at the user or organization level.

Use cases

MCP servers on Cloudflare \ \Deploy MCP servers on Cloudflare Workers with Stytch handling authentication and authorization.

MCP servers on Vercel \ \Build MCP servers on Vercel’s edge runtime with secure token management.

CLI applications \ \Authenticate command-line tools using the Authorization Code Flow with PKCE for public clients.

Third-party integrations \ \Allow external developers to build apps that integrate with your platform through OAuth.

Start building

Connected Apps overview \ \Learn the fundamentals of OAuth, OIDC, and how Connected Apps works.

API reference \ \Complete API documentation for token exchange, consent management, and client configuration.