#Deploy

Ship a broken release on purpose, watch your users stay on the old version, then roll back with one command. Real commands, real output, self-hosted.

A postmortem from our own dogfood run: a failed readiness probe still received traffic. What broke, why two correct parts added up to a bug, and the fix.

TL;DR: docker compose up is a development command; there is no compose-to-cloud button, and service names like http://api:8080 only resolve on the machine running Compose.…

Deploy apps to Vercel without an account using Vercel CLI. Share prototypes on tmp.now, then sign in to claim your project within one hour to keep it live.

How a conference talk, a hardcoded branch, and too many deployment requests led me to ephemeral environments.

I run coding agents the way most of us do now: it proposes, I approve, it executes. Last week I...

UPDATE 22:25 UTC: Google appears to have fixed it. Cold starts are working again (seen i n us-central1) even for functions still on automatic updates. The on-deploy redeploy below is still a valid fallback if it comes back.…

Ship changes without downtime or big-bang risk. Recreate, rolling, blue-green, canary, and progressive delivery with automated analysis and rollback,

I recently finished the Cloud Resume Challenge: host a resume as a static site on a cloud provider,...

The pipeline is green and production serves the old commit. How this happens, a script that catches it, and the headers and ordering that fix it.

Originally published on the Hubfly space blog. TL;DR What you get: every push to main redeploys...

Forge app CI/CD in Bitbucket Pipelines: the Jira Coder writes a UI Kit issue panel, the pipeline locks its scopes and runs forge deploy under secured variables.

Adding AI models to local applications requires a portable model format, a reliable runtime, and acceleration that works across target systems.

The Next.js AI Chatbot template has been updated to use AI SDK 3.0 with React Server Components, including Generative UI examples.

You can now deploy FastHTML Python projects on Vercel with zero configuration, with Vercel CLI or by pushing new changes to your git repository.

Vercel Drop lets you deploy a file, folder, or .zip by dragging it into your browser. No Git or CLI required. Framework projects build automatically.

We've removed CLI-specific deployment limits, making it easier to deploy from local machine, agent workflows, and external CI/CD pipelines with instant feedback.

You can now inspect the framework preset and files Vercel CLI will include in a deployment—without uploading files or creating a deployment.

The Vercel MCP server can now deploy code directly. Send files with the deploy_to_vercel tool to create a project, build, and get a URL, with no git repo or CLI.

Use the Vercel CLI to deploy static artifacts under one second. When Vercel detects a valid static artifact, it skips the build step and returns a live URL instantly.