Fetching from the wire…
Public story · 2026-07-25 · high
Hobby stays capped at 300 seconds, but paid tiers can run a Workflow step for 30 minutes with one env flag and a redeploy.
Why now: This is the Workflow-duration change Vercel's changelog surfaced in the July 25 coverage, still marked beta with no graduation date given.
Vercel raised the ceiling on Workflow step runtime from roughly 800 seconds to 1,800, per its changelog. It's a beta feature, gated behind an environment variable called VERCEL_ENABLE_WORKFLOW_EXTENDED_MAX_DURATION=1, and it needs a redeploy plus Fluid compute and a compatible Node.js or Python runtime to take effect. Hobby projects don't get it. They stay capped at 300 seconds.
This matters most if you've built anything with a long-running agent step, a batch job, or a multi-stage generation task that used to blow past 13 minutes. The old limit forced a specific pattern: break the work into chunks, checkpoint state somewhere, and resume on the next invocation. That's real code, real state management, and real edge cases around partial failures.
With a 30-minute ceiling, a lot of that scaffolding becomes unnecessary. If your longest step ran 14 to 25 minutes, you can likely collapse it back into one function call and delete the resume logic entirely.
The catch is the runtime requirement. This isn't a flip-a-switch change on every project. You need Fluid compute enabled and a Node.js or Python runtime that supports the extended duration, so teams on older runtime pins or non-Fluid deployments don't get the benefit until they migrate. The changelog doesn't say when this graduates out of beta or whether 1,800 seconds is a hard ceiling or a placeholder.
Worth watching: whether Vercel treats this as a stepping stone toward removing step-duration limits entirely, or whether 30 minutes is where they intend to stop. For anyone running agent pipelines with genuinely open-ended steps, that distinction decides whether you drop the resume mechanism entirely or only loosen its trigger threshold.
Each link below shares sources, entities, or timing with this story.
Vercel just shipped the stuff that makes Fly.io and Railway nervous, and they did it in a single push. The Vercel Container Registry is now a thing: an OCI-compliant image registry on Vercel's own infrastructure. You can deploy an HTTP server straight from a Dockerfile or Cont...
Python functions now compile to bytecode during build rather than at first import, a 54% median cold-start reduction with zero code changes (Vercel). Python cold start has been the main reason teams reach for Node on edge functions. This doesn't close the gap but it narrows it...
He handed Claude Fable 5, in Claude Code for web, the job of evaluating whether smolvm 1.8.3 could safely run untrusted Python and JavaScript under strict constraints. Verdict: it fits, with hardware-isolated VMs rather than shared-kernel containers, enforced CPU/RAM limits, g...
virgiliojr94/book-to-skill leads GitHub weekly Python trending for 16,749 total (1,781 forks, MIT, created May 1). It converts a technical book PDF into a Claude Code skill directory agents reference while working. Nobody in the repo addresses the copyright surface of bulk-con...
Vizhi (mission control for Codex CLI on a Logitech MX keypad), AgentLoop (spawns a fresh Codex worker and critic every cycle), and AgentManager (alerts when a Claude Code session is waiting for input) all launched together (Product Hunt). Physical controls for agent supervisio...
This is the most useful thing I read this week and it isn't close. Anthropic published its internal methodology for running large-scale code migrations with Claude Code on July 16, and unlike most engineering-blog playbooks, it carries receipts. Bun's Zig→Rust migration: rough...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.