GitLab 19.0 adds AI workflows, secrets management, and self-hosted model support

GitLab has rolled out version 19.0 of its platform, delivering a suite of new features designed to streamline development, bolster security, and enhance visibility for platform engineering teams. A major focus of this release is the integration of AI workflows, alongside robust secrets management and improved CI pipeline insights. One of the standout additions is the GitLab Secrets Manager, now in public beta for Premium and Ultimate users. This feature allows credentials to be stored securely within the same platform that executes code and pipelines. Secrets are scoped to only authorized jobs, and access control and audit logging leverage existing GitLab group and project structures, eliminating the need to maintain separate permission models. In the event of a credential compromise, responders can trace its usage directly from the GitLab audit trail, linked to the originating pipeline, without having to correlate logs across disparate systems. Furthermore, GitLab 19.0 addresses the growing "AI Paradox" where AI accelerates code generation but not necessarily its trustworthiness or security at scale. The platform's "agentic core" is advanced by embedding capabilities directly where teams work, aiming to reduce handoffs between code writing and shipping. This includes agentic merge request workflows that extend across the full MR lifecycle, assisting with reviewer feedback, conflict resolution, and splitting oversized merge requests. For platform engineering teams, Components Analytics offers crucial visibility into shared CI infrastructure. This feature provides insights into which CI/CD Catalog components are in use across an organization and their respective versions. All this data resides within GitLab's unified platform, enabling teams to act on information without switching tools. Adoption data is available for all tiers, with per-component drill-down for Ultimate users. The release also strengthens software supply chain visibility with new security capabilities. Dependency scanning, now with a software bill of materials (SBOM), creates an auditable inventory of third-party components, matched against GitLab security advisories. This gives Ultimate tier users clear evidence of what entered each build. Security configuration profiles further simplify governance, allowing teams to enable features like Secret Detection, SAST, and Dependency Scanning through policies rather than individual project CI configurations. These comprehensive updates underscore GitLab's commitment to providing an integrated platform that supports rapid, secure, and governed software delivery in the age of AI.