GitLab's Developer Flow now automates merge requests end to end

In GitLab 19.0, the Developer Flow AI agent no longer stops at writing code. It now handles the entire merge request lifecycle, from addressing reviewer comments and resolving merge conflicts to the final merge. This is the next step after last year's launch of Developer Flow, which at the time could only turn an issue into a ready MR.

What changed in GitLab 19.0

According to the GitLab Blog, AI had until now only sped up writing code, while everything between opening a merge request and merging it stayed manual: assigning reviewers, handling feedback round after round, untangling conflicts, rebasing before the merge. Every step demanded the developer's attention. The bottleneck simply moved, and the tools didn't adapt to it.

Developer Flow launched last year and solved just one problem: it turned an issue into a merge request, automating the routine prep work between defining a task and getting an MR up for review. But once the MR existed, all further iteration on it remained fully manual.

The new version of Developer Flow closes exactly that gap. The same AI agent now handles reviewer feedback, fixes conflicts in long-lived branches, works its way through unfamiliar code, and splits oversized MRs into smaller pieces. It pairs with autonomous merge conflict resolution and one-click rebase and merge.

How the agent joins work on an MR

You can launch Developer Flow at any stage of work on a task. There are three entry points: the Generate MR button directly on an issue, assigning the Duo Developer service account directly to an issue or MR, or a new @-mention trigger in any discussion thread on an issue or merge request.

The key difference from the previous version is that the agent picks up the conversation and keeps working on the same MR, instead of creating a new one that someone would then have to review manually from scratch. Combined with autonomous merge conflict resolution and one-click rebase and merge, this cuts the amount of manual work across the whole stretch between opening an MR and its final merge.

A new category of AI development tools

GitLab describes Developer Flow as part of a new, still-emerging category of AI coding tools. The first wave of these tools sped up writing the next line of code (autocomplete); the second gave developers a chat window for questions. What is emerging now is fundamentally different: agents that take part in the work from start to finish, not at a single fixed moment.

In GitLab's framing, developers can now stay "above the loop": delegating tasks and reviewing results while the agent handles the hands-on execution — reviews, conflicts, refactoring MRs that have grown too large, and exploring unfamiliar parts of the codebase.

What is Developer Flow in GitLab 19.0?

Developer Flow is GitLab's AI agent. In version 19.0, its scope expanded from creating a merge request to covering the request's entire lifecycle: handling reviewer feedback, resolving merge conflicts on long-lived branches, exploring unfamiliar code, and splitting oversized MRs into smaller pieces. It works alongside autonomous conflict resolution and one-click rebase and merge.

How does the new

Developer Flow differ from last year's version?

Last year's Developer Flow launch automated only one stage: turning an issue into a merge request, meaning the routine prep work between defining a task and getting an MR up for review. Any further work on an existing MR remained fully manual. In GitLab 19.0, the same agent keeps working on the same MR after it is created: it addresses comments, fixes conflicts, and helps get it ready to merge.

How do you run Developer Flow on a merge request?

There are three ways: click the Generate MR button on an issue, assign the Duo Developer service account directly to an issue or merge request, or mention the agent with an @-trigger in any discussion thread on an issue or MR. In every case, the agent picks up the context of the discussion and continues working on the existing MR rather than creating a new one.