Docker, Inc

Claude Opus 4.7 is now supported in Docker Agent. Docker Agent is a framework for building + running custom agent teams. We support the models powering your work today, and we’ll keep adding the ones you’ll reach for next. Check out the latest release → https://bit.ly/42f0nw8

📣The Call for Speakers is extended for WeAreDevelopers World Congress North America! AI is moving fast, and the best lessons are coming from people shipping it for real. WeAreDevelopers North America is a new builders-first event for the AI era: hands-on, workflow-driven, and focused on what it takes to ship AI systems safely and at scale. With Docker co-hosting, we’re bringing together 10,000+ developers and practitioners from hundreds of leading companies to build, test, and share real patterns across agentic development, platform engineering, security, and cloud-native delivery. If you have a talk, demo, or hard-earned lesson to share, this is your opportunity to take the stage. 📍 San Jose • Sept 23–25 📝 CFP extended through April 30 Submit your talk → https://lnkd.in/dkhECnfc

Most sandboxing approaches force a tradeoff. Speed vs isolation. Flexibility vs security. Developer experience vs control. This post breaks down why Docker took a different path with Sandboxes - using microVMs to give agents a real development environment with strong isolation, without adding friction. It walks through: • Why containers and VMs fall short for agent workflows • How microVMs enable full Docker environments with no host access • Why we built a new cross-platform VMM from scratch • What this means for running agents safely in practice If you’re thinking about how to run agents autonomously without compromising your system, this is a useful look at the architecture behind it. Read more →

Running AI is no longer just about models. It’s about execution, isolation, and control. This issue of Docker Navigator explores what that looks like in practice: • Sandboxes for autonomous agents • Local model workflows • Running Docker anywhere with Offload • Supply chain security and reliability A practical look at what it takes to move from experimentation to production. Read the full newsletter →

Here is a great example of what it looks like to take Sandboxes from concept to real-world use. This walkthrough puts the isolation model to the test - probing boundaries, validating how it holds up, and digging into how credential injection and autonomous agent workflows behave under real conditions. This is a great read if you want a clearer understanding of what’s actually happening under the hood and how to apply these patterns in your own workflows.

Many Hugging Face Spaces aren’t built with Arm64 in mind, and issues aren’t always obvious until you try to run them. Using a recent real-world example (ACE-Step v1.5), this guide shows how to scan a Space for Arm64 readiness in minutes, not hours. With Docker MCP Toolkit + Arm MCP Server, you can: - Identify architecture blockers before you build - Surface hidden dependency issues (like x86-only wheels) - Get a clear GO / NO-GO migration verdict If you’re targeting Apple Silicon, Arm cloud, or edge deployments, this is a practical way to validate compatibility early. Read the full walkthrough → 

Want to run Hugging Face Spaces on Arm64 without trial and error? Join us tomorrow for a live code-along: Docker + Arm walk through how to scan a Space, surface compatibility gaps, and apply fixes using Docker MCP Toolkit, the Arm MCP Server, and GitHub Copilot in VS Code. You’ll see: • How to assess Arm64 readiness • How MCP servers analyze containers, code, and dependencies • How to catch and fix architecture-specific issues • A full demo using a real model (ACE-Step v1.5) If you’re working with Apple Silicon, Arm cloud, or edge AI, this is a practical walkthrough you can follow along with. 📅 April 16 🕘 12:00pm EST / 9:00am PST / 9:30pm IST Join live → https://lnkd.in/gU2k2GQW

New episode of Ship Happens is out. Per Ploug Krogslund sits down with WordPress VIP CTO Brian Alvey to discuss how the web has evolved: from the early days of blogging to today’s AI-driven, platform-centric landscape. They explore how AI is reshaping the economics of software development and what happens when bots and agents consume content without attribution. They also get into why open source continues to play a foundational role in innovation, especially as large language models rely on open data to operate and improve. Tune in for a practical look at where the web is heading, and what it means for developers building in an AI-first world. Catch the full conversation → https://lnkd.in/geywr7DJ

Building an AI agent is one thing. Getting it into production… that’s where things get real. In Episode 4 of Global AI Community’s Made for Dev series, Oleg Šelajev and Sammy Deprez walk through how to go from local scripts → production-ready AI services using Docker. If you’re moving beyond prototypes, this is the part that matters. 

Docker Captain Kristiyan Velkov's production-ready examples for containerizing Next.js apps are now part of the official Next.js repository! This is what great developer experience looks like: → Real-world, scalable patterns → Built from lessons across React, Node, Angular, Vue → Designed for production, not just demos Shoutout to Kristiyan, and to the Vercel + Docker, Inc teams for continuing to support the ecosystem and make it easier for developers to go from local to production. Check it out: https://bit.ly/4te03cX