GitLab

Agentic Memory and GitLab 19.0

Choosing your model is the starting line, not the finish. GitHub's BYOK and local model support in Copilot CLI gives developers more control over where their code and prompts go, which is progress. What it doesn't change is the bigger problem one layer deeper. When AI starts triggering builds and acting across your pipeline, true enterprise governance is what determines whether those actions are safe, controlled, and auditable. GitLab Duo CLI is built for exactly that. Here’s our take on what enterprise-grade AI automation requires. https://lnkd.in/g-4963sR

AI is generating more code than ever, and the work surrounding that code hasn't kept up. GitLab 19.0 helps close that gap by extending agentic AI and platform controls across the full software delivery lifecycle. ✨ What's new: ➡️ GitLab Secrets Manager (Public Beta) keeps credentials in the same platform that runs your code, scoped per job and governed by the access controls you already use. ➡️ Developer Flow now extends across the full MR lifecycle, paired with autonomous merge conflict resolution (Beta) and one-click rebase-and-merge (Beta) to cut manual work between opening an MR and merging it. ➡️ Components Analytics for the CI/CD Catalog gives component maintainers usage data across all tiers. Drill into any component to see which projects are on which versions, and where security fixes still haven't landed. ➡️Expanded model support for Duo Agent Platform Self-Hosted adds four new open source models, giving air-gapped and regulated environments more capable options for agentic workflows. …and more! Read the full 19.0 release notes: https://lnkd.in/gyySGrX2 🚀

GitLab Secrets Manager is now in public beta. It keeps credentials in the same platform that runs your code and pipelines, with no standalone vault to maintain. Each secret is scoped to the job that needs it, and governed by the access controls you already use for code. If a credential is exposed, the blast radius stays small. Watch the demo to learn more. https://lnkd.in/gNhd8nhW

Stop manually untangling merge conflicts. GitLab 19.0's Developer Flow reads both branches, picks a resolution strategy, commits, and explains what it did. See it in action below. Learn more: https://lnkd.in/g2GS5Hzm

Pipeline standards drift the moment they ship. Components get copy-pasted, versions fall behind, and you lose track of what's actually running. GitLab 19.0 is here to help. Components Analytics for the CI/CD Catalog introduces two new views that show how your shared pipeline components are being used across your organization. Component maintainers can see usage frequency across all tiers. With GitLab Ultimate, drill into any component to see exactly which projects are running which versions. Watch the demo. https://lnkd.in/gtR9p69f

Air-gapped and regulated environments have historically been a step behind on AI capability, since the most capable models tend to land in cloud-first deployments. GitLab 19.0 narrows that gap. Learn more. https://lnkd.in/gNdgM4_S GitLab Duo Agent Platform Self-Hosted now supports four new open source models, Mistral Devstral 2 123B, GLM-5.1, Kimi-K2.6, and MiniMax-M2.7, so teams can match the right model to the right workflow. Inference runs on your hardware, and no data leaves your environment.

AI made writing code dramatically faster, but the work between opening an MR and merging it has stayed almost entirely manual. GitLab 19.0 changes that. Developer Flow now extends across the full MR lifecycle, with a single agent that addresses reviewer feedback across multiple rounds, resolves conflicts on long-running branches, researches unfamiliar codebases, and splits MRs that grew too large. This enables you to stay above the loop, steering, reviewing, and deciding while your agent handles execution. Watch the demo to learn more. https://lnkd.in/g2GS5Hzm

AI is generating more code than ever, and the work surrounding that code needs to keep up. GitLab 19.0 expands our Intelligent Orchestration to close the gap between writing code and shipping it. Highlights include: 🟠 GitLab Secrets Manager, now in public beta, scopes each credential to the job that needs it and governs access through the same controls used for code. 🟠Developer Flow now handles the full merge request lifecycle, from reviewer feedback and conflict resolution to one-click rebase-and-merge. 🟠Components Analytics gives platform engineering teams visibility into which CI/CD Catalog components and versions are running across their organization. 🟠GitLab Duo Agent Platform Self-Hosted gains four new open source model options for teams using air-gapped and regulated environments. 🟠Dependency scanning with a SBOM and security configuration profiles gives teams auditable control and visibility over what ships. Learn more about the 19.0 release: https://lnkd.in/gJTgMFB9 Ready to dive deeper into 19.0? The full release blogs are linked in the comments.

From terminal to merged MR, with the agent doing more of the heavy lifting at each step. All thanks to Codex + GitLab. In this new tutorial, you'll learn how to use Codex and GitLab Duo Agent Platform to: 1. Fix a Rust WebSocket bug locally to get started with Codex 2. Enrich the context with GitLab MCP to fix the Rust bug aligned to issue requirements 3. Use Codex within GitLab Duo Agent Platform as an external agent to help address review feedback in the merge request Read the full tutorial: https://lnkd.in/g_TNjzCa