GitLab has made its Duo Agent Platform generally available. With this general availability, the company is taking the next step in integrating agentic AI into the entire DevOps process.
The platform is designed to support development teams with AI-driven automation that works within an organization’s existing context, standards, and governance.
Central to the GA release is Agentic Chat. This functionality builds on the previously introduced Duo Chat but goes a step further by leveraging context from virtually every part of GitLab. Think of issues, merge requests, CI/CD pipelines, and security findings. Agentic Chat can not only advise, but also actually perform actions on behalf of developers, depending on the rights and approvals that have been set.
Platform as part of daily work
Within the GitLab web interface, Agentic Chat can create issues, epics, and merge requests, among other things, and summarize relevant findings into concrete follow-up steps. In development environments such as Visual Studio Code and JetBrains IDEs, the chat helps generate code, configurations, and infrastructure-as-code. In addition, it can fix bugs, write documentation, and produce supporting texts. GitLab explicitly positions this functionality as part of developers’ daily work, rather than as a separate AI assistant.
The Duo Agent Platform also provides pre-built foundational agents ready for use. Two are available at launch. The Planner agent helps structure and prioritize work and split tasks within GitLab. The Security Analyst agent focuses on analyzing vulnerabilities and other security signals and explains the impact and possible next steps. GitLab indicates that additional agents are currently in beta and will be available later.
In addition to these standard agents, organizations can also build their own agents, tailored to internal processes and policies. The platform also supports the use of external agents and tools, such as Claude Code or OpenAI Codex CLI. GitLab can be linked to other systems, including Jira, Slack, Confluence, Playwright, and Grafana, via an MCP Client. In this way, GitLab aims to offer teams flexibility without losing central control over data and security.
Agentic Flows for complex tasks
Another important part of the platform is Agentic Flows. These allow multiple agents to be used in conjunction to perform more complex tasks, such as setting up a merge request from an issue, migrating or repairing CI/CD pipelines, or streamlining code reviews. These flows combine automation with control, whereby human approval can remain part of the process.
In terms of governance, GitLab focuses explicitly on enterprise environments. The platform is available on GitLab.com and in self-managed installations from the 18.8 release onwards. Organizations gain control over model selection per namespace, group-based access rights, and approval flows for tool usage. Integrations with LDAP and SAML ensure that the use of agentic AI aligns with existing identity and security structures. With this, GitLab aims to embed AI deeply in the development process without compromising compliance and management.