Blog

Claude Code speaks Anthropic's Messages protocol and so do the major gateways. A small shell wrapper points it at any model, with live rate cards to prove it.

WorkOS is becoming an AI-native organization. The Applied AI Showcase in NYC on July 28th 2026 showed exactly how with live demos and talks from the Applied AI team.

How SCIM Bridge moves live enterprise directories onto WorkOS Directory Sync with zero downtime, staged authority, and lossless rollback until the last step.

Wiring AuthKit through Electron's main, preload, and renderer processes, without shipping a secret in your binary.

We built audit logging for Claude Code, Codex, OpenClaw and pi with no API key on any laptop. Use it to get total visibility into your entire org's agentic activity.

A runnable tutorial for pulling per-connection third-party credentials from WorkOS Pipes with nobody logged in, plus relay, rotation, revocation, and audit logging.

Dynamic Client Registration is deprecated, and Client ID Metadata Documents answer only half the question. Here is what each pattern actually proves, and how auth.md adds the user delegation they leave out.

Your agent calls a third-party API for a user who never connected it. What to send back, why 401 and 403 are both already taken, and what the x402 standard means for the answer.

Build a Next.js app with AuthKit sign-in, then gate a destructive action behind a fresh re-authentication using auth_time and max_age.

Agents that call third-party APIs on behalf of users normally hold the OAuth access token. Here is where that token leaks, and how to get delegated access without it.

A step-by-step guide to PKCE-based sign-in with the WorkOS iOS SDK's PublicClient.

In AI agent governance, most MCP servers run as local processes that never cross the network boundary older shadow-IT tools were built to watch. The risk changed too: from data leaving through an unsanctioned app to actions taken through one.

A recap of the August 5 Demo Night at our San Francisco office: 21 live five-minute demos, no slides, no pitches, and an agent harness running on Mac OS 9.

In AI agent governance, the analogy holds for access and response speed but breaks on motive. Google DeepMind's review of a million agent trajectories found most incidents come from overeagerness, not malice.

Town co-founder Jean-Denis Greze talks to Michael Grinich at AI Engineer World's Fair 2026 about draft-only agents, earning trust, and shared data silos.