About GitLab
GitLab states that software teams and their agents plan, build, secure, and ship on one control plane. GitLab describes that control plane as covering unified context, orchestration, security and governance, and execution on GitLab Core. GitLab combines planning, source control, and CI/CD in one platform, and projects, releases, and code share one data plane.[3][4]
GitLab runs four peer pillars on one foundation. Execution and Workflows covers planning, source code management, CI/CD pipelines, registry, and delivery. Security and Governance covers application security, software composition analysis, compliance management, and agent governance. Context and Data covers analytics, value stream visibility, and lifecycle context including Context Graph. Agentic Orchestration covers GitLab Duo Agent Platform: agents, agentic workflows, model choice, and extensibility. Deployment and AI inference options let customers choose where GitLab runs and which models power agents.[2][3]
GitLab Duo Agent Platform is how GitLab orchestrates agents across the software development lifecycle. GitLab says humans stay above the loop with policy, identity, and audit. GitLab defines intelligent orchestration as teams setting intent and guardrails while agents execute within them. Policies, approvals, and scanning in GitLab apply to human and agent changes alike. GitLab supports APIs, webhooks, integrations, and Model Context Protocol (MCP) so agents and tools share context under the customer's policies. Across deployments, GitLab customers can use GitLab-managed models or self-hosted models, and reshape seats and agent usage with GitLab Flex.[3][4]
GitLab runs scanning, policies, and approvals inside the pipeline. GitLab consolidates scanners such as SAST, SCA, Secret Detection, and DAST into one platform. Security findings from GitLab appear directly in merge requests and IDEs. GitLab built-in scans run on every push to detect insecure code during development. GitLab can run automated security scans to capture vulnerabilities before code is merged. GitLab supports ensuring that passwords and sensitive information are not in source code. GitLab can find disclosed vulnerabilities in dependent software before they reach production. GitLab can generate an SBOM automatically so that every dependency in the project is known.[4][5][6]
GitLab can block risky packages before they enter the build. GitLab describes protecting multiple attack surfaces, including code, build, dependencies, and release artifacts, with policy-as-code and continuous governance. GitLab can block malicious packages and packages above a vulnerability severity threshold before they are fetched. GitLab can scan licenses and block packages that fall outside organization policies. GitLab isolates the build environment to block unauthorized access and malicious code execution.[6]
GitLab offers over 60 out-of-the-box controls prebuilt into a single platform. GitLab keeps code authors, reviewers, and deployers separated by policy so that no single person or agent can develop and deploy unreviewed changes. GitLab captures every human and AI-driven change in one immutable audit trail. GitLab maps controls from every major or custom framework into a single, unified compliance view. For infrastructure, GitLab includes Infrastructure as Code static and dynamic testing to help catch vulnerabilities before they get to production. GitLab integrates Flux to support pull-based GitOps workflows. GitLab stores Terraform state files and modules and shows Terraform plan output directly in the merge request.[4][7][8]
GitLab.com is multi-tenant SaaS, and Self-Managed runs in the customer's environment on any cloud or on premises, including air-gapped patterns. GitLab Dedicated and GitLab Dedicated for Government provide single-tenant SaaS with data residency and isolation controls. GitLab can deploy to virtual machines, Kubernetes clusters, or FaaS from multiple cloud vendors. Organizations can make one annual commitment and steer their spend across seat-based licenses and usage-based credits.[3][4][9]
GitLab provides building blocks to support compliance, certification, and standards programs for functional safety, code quality, and other requirements: ISO 26262, MISRA, IEC 61508. GitLab describes integrating virtual and physical hardware testing into automated CI. GitLab describes self-hosted and BYOM options that keep code and data inside the customer's boundary. GitLab describes shift-left policies, remediation, and release evidence that help engineer processes for DO-178C.[2][3][9][10][11][12][13]
GitLab on the record