Docker, Inc

Docker Model Runner now supports Vulkan! This unlocks hardware acceleration for LLMs on way more machines - including integrated GPUs and GPUs from AMD, Intel, and others. It’s a big step toward democratizing local AI: - No dependence on NVIDIA or Apple Silicon - Model Runner detects Vulkan and uses your GPU automatically - Just run 'docker model run ai/gemma3' More hardware. More speed. More local AI. #Docker #ModelRunner #LLM #AIInfra #Vulkan #GPU #OpenSourceAI #LocalAI #DevTools

What happens when agentic AI meets Java? Spring Framework creator Rod Johnson joins Oleg Šelajev in the latest episode of AI Guide to the Galaxy to explore Embabel, an agentic AI framework for the JVM built with Spring and Spring AI. Learn how Embabel brings AI agents into enterprise environments through: • Deterministic planning for reliable multi-step workflows • Strong typing and type-safe prompts • Real-world demos in Java and Kotlin using the Docker Desktop MCP gateway See why some AI logic belongs in code, not in prompts, and how developers can build explainable, production-grade agentic systems on familiar foundations. 📺 Watch the full episode: https://lnkd.in/e4d-TvwF #AIGuideToTheGalaxy #Docker #SpringAI #AgenticAI #JavaDevelopers

Check IBM's latest family of open-source language models, Granite 4.0, on Docker Hub! This latest release delivers open, high-performance SLMs with faster inference, long-context support, and easier customization - designed for real-world enterprise AI. What’s new: - Hybrid MoE + Mamba-2 architecture for speed and efficiency - Up to 128k context lengths - Run on GPUs as small as a 3060 - Apache 2.0 licensed, ISO 42001 certified - Multiple model sizes from 3B to 32B - Built for RAG, agents, and edge AI #Granite #OpenSourceAI #LLMs #IBM #SLMs #DevTools #GenAI #EnterpriseAI #RAG #AgenticAI

Shipping secure software shouldn’t be a luxury. Starting today, with an add-on subscription, you get unlimited access to the Docker Hardened Images catalog - so every team can start from a secure base, without the premium price tag. Built from source - Continuously patched - Up to 95% smaller - VEX support included - Independently validated by SRLabs From Python to PostgreSQL to FedRAMP-ready variants, this is a new foundation for building secure-by-default. Read the blog and start a 30-day free trial (no hoops, just Docker ID): #Docker #ContainerSecurity #SupplyChainSecurity #OpenSourceSecurity #DevSecOps #SoftwareSecurity #CVE

Open Source Genius had a familiar dev problem: 15–30 docs hitting downloads folders daily across teams. Then they built an agent using Docker’s MCP Toolkit + Claude Desktop . - Fully automated workflow - 2 hours saved per dev every week - No cloud, no SaaS, no new approvals - Built in under an hour. Read how they did it: #Docker #DevTools #MCP #Claude #Automation #DeveloperExperience

Going to Devoxx Belgium? Make time for this session. On Wednesday, Oct 8, Docker’s Guillaume LOURS is going live - no fluff, just a fast-paced demo of how Docker Compose is evolving for AI and cloud-native dev workflows. You'll learn how to: - Define local or remote LLMs directly in your docker-compose.yaml - Use provider services to prep external resources like inference endpoints - Seamlessly offload GPU-heavy tasks to the cloud with Docker Offload - Treat your Compose file as the single source of truth for modern, AI-powered apps Whether you're building with LLMs, agents, or distributed services, this session is packed with practical takeaways for streamlining dev with Compose. More info: https://lnkd.in/e9CmSaqG #Docker #DockerCompose #Devoxx #AIInfra #LLM #CloudNative #AIWorkflow

What did we learn at AI Engineer Paris? That the future of software is agentic - and developers want better tools to build it. With Docker, building, sharing, and securing agents is as simple and portable as containerizing microservices. In our keynote, workshop, and live demos, we showed how Docker helps devs: - Build collaborative AI agents with cagent - Package & share agents like containers - Run MCP servers with zero config Missed it? Catch the recap blog + workshop: #Docker #AIEngineerParis #AgenticAI #DevTools #MCP #ModelRunner #cagent #AIInfrastructure

What if building AI agents felt like writing a docker-compose.yaml? In Episode 5 of AI Guide to the Galaxy, Oleg Šelajev and Djordje Lukic demo cagent, Docker’s open-source agent builder that makes it dead simple to: - Define agents in YAML - Run them locally or in the cloud - Plug into MCP servers and share on Docker Hub This is agentic AI, the Docker way. 📹Watch now: https://lnkd.in/g8MrVbZZ And stay tuned, the next episode is dropping soon. #AIGuideToTheGalaxy #Docker #cagent #AgenticAI

llama.cpp just got smarter: Resumable GGUF model downloads are now live - no more starting over on network hiccups. Coming soon: Docker Model Runner will bring this same reliability to a Docker-native AI workflow. #Docker #ModelRunner #llamacpp #MLOps #LLM #DevTools #AIInfrastructure #OpenSourceAI

Check out the next two weeks of our Docker Captains’ schedule! Our Captains are all over the world, and we have a great lineup 👇