GitLab

Expanding AI usage across your org is harder to justify without spend transparency and controls. GitLab now offers subscription-level and per-user spending caps for GitLab Credits. This gives teams cost predictability, enforcement controls, and a path to deploying GitLab Duo Agent Platform at scale. You get the flexibility of usage-based consumption paired with the cost controls that finance needs to approve the rollout. Learn more. Link in comments.

AI-generated code moves faster than the systems around it can keep up with. In 18.11, GitLab ships two agents built to address the gaps a general-purpose assistant can't. The CI Expert Agent (Beta) inspects your repo and proposes a working pipeline in minutes. The Data Analyst Agent (GA) answers natural-language questions about your SDLC data directly in Agentic Chat, no query language or dashboard request required. Both run inside the platform where your code, pipelines, and issues already live, providing context that no external tool can match. Learn more. Link in comments.

Faster code generation hasn't meant faster delivery. GitLab 18.11 helps address that gap with agentic AI across pipelines, security, and delivery analytics. ✨ What's new: ➡️ Agentic SAST Vulnerability Resolution is now GA, automatically generating ready-to-merge code fixes to close vulnerabilities before they reach production. ➡️ Two new foundational agents join GitLab Duo Agent Platform: the CI Expert Agent (beta) targets a working pipeline in minutes, and the Data Analyst Agent (GA) surfaces live delivery insights in natural language. ➡️ New subscription-level and per-user spending caps for GitLab Credits give organizations control over on-demand AI spend. …and more! Read the full 18.11 release notes: https://lnkd.in/geZP55Mr 🚀

AI-generated code moves faster than the systems around it can keep up with, creating the AI Paradox. As code volume grows, so does the backlog of pipelines to configure, security findings to remediate, and delivery questions to answer. GitLab 18.11 helps address those gaps by applying agentic AI to workflows that come after the code is created, from security remediation to pipeline setup. Highlights include: ⭐ Agentic SAST Vulnerability Resolution (GA) is now available to Ultimate customers, automatically generating ready-to-merge code fixes so teams catch vulnerabilities before they reach production. ⭐ Two new foundational agents: the CI Expert Agent (beta) can help stand up a working CI pipeline in minutes with no manual YAML, and the Data Analyst Agent (GA) answers natural-language questions about merge request cycle times, pipeline health, deployment frequency, and more. ⭐ Usage Controls: New subscription-level and per-user spending caps for GitLab Credits give organizations cost predictability when deploying GitLab Duo Agent Platform at scale. Learn more about the 18.11 release: https://lnkd.in/gka_mTRB Ready to dive deeper into 18.11? The full release blogs are linked in the comments.

Claude Opus 4.7 from Anthropic is now available in GitLab Duo Agent Platform. Available now via model selection. https://lnkd.in/g_vn67ri Teams running agents across code review, CI/CD, and security remediation get stronger reasoning and more precise instruction-following.

Everyone's focused on building AI agents. Fewer are asking what comes next. Once it's built, you still need to secure it and deploy it to the right environment. Before heading to Las Vegas for Google Cloud Next, our own Regnard Raquedan, Sr. Ecosystem Solutions Architect at GitLab, previews his talk about how GitLab and Google Cloud Agent Engine bring security and deployment speed together for agentic development. Catch his lightning talk, "Secure and Fast Deployments to Agent Engine with GitLab," at Google Cloud Next, April 22-24, booth #2324.

In the latest episode of The Developer Show, we answered community questions about how and when to use multi-agents. 👇

The GitLab Duo CLI is now in beta, bringing the full power of GitLab Duo Agent Platform directly to your terminal. Developer Advocate Daniel Helfand walks through how to get started with Duo CLI for local development workflows. You'll see authentication set up, pulling context from GitLab issues and Agentic Chat sessions, and taking a feature from plan mode through merge request creation and deployment. Learn more in the link in the comments.

GitLab Duo Agent Platform agents can now call foundation models through Vertex AI, including Gemini. AI agents that build software need more than a capable model. They need context from the issue that started the work, the repository it touches, the pipeline that validates the result, and the security policies that govern what ships. GitLab is where that context lives. Every agent action flows through the same access controls, approval rules, and audit logging that govern human developers. Governance does not stop when an agent takes over. Joint customers can run GitLab's AI Gateway on Google Cloud infrastructure and count GitLab Duo Agent Platform usage toward existing Google Cloud commitments. Read more: https://lnkd.in/gDmZXZwf

AI-powered test insights are only as good as the data behind them. Connecting GitLab pipelines to SmartBear QMetry gives its AI engine real-time data to analyze test execution patterns, catch flaky tests, predict failures before they surface, and recommend exactly where to optimize. The more continuous the data, the sharper the insights, and the more your team can focus on work that actually moves the needle.