GitKraken Unveils Code Flow to Help Teams Navigate the AI Era
With AI adoption accelerating and the nature of coding fundamentally shifting, GitKraken has introduced Code Flow, a framework for describing how work flows between developers, coding agents, repositories, reviews, pull requests, and ultimately the production environment. This, because AI has increasingly moved developers away from writing every line of code themselves and toward a role where they manage, coordinate, and review the output of multiple, simultaneous AI agents.
This new way of working creates a complex environment. Historically, development tools were built to track code written by humans. Now, developers must handle a much larger volume of work, including more branches, pull requests, and commits, all generated by agents across multiple repositories. As GitKraken CEO Matt Johnston noted, the difficulty is no longer about generating code—it is about turning all that agent-generated work into high-quality, finished software.
“Eleven years of watching how code moves through real teams taught us exactly where things break down,” he said. “But what we are seeing now is different in scale. AI has multiplied the volume of code flowing through those same failure points. The review that never gets picked up is now three reviews. The merge conflict that becomes a blocker is now five. The branch that drifted so far from main that the developer who opened it is afraid to touch it, that used to happen occasionally. Now it happens every sprint.”
This is where the concept of Code Flow comes in. Code Flow is designed to describe the entire life cycle of software development in an agent-heavy environment. It tracks how work moves through every stage: from the initial planning and coding to reviews, branching, and eventually, the production environment.
Who is Code Flow for?
As Johnston explained to SD Times: “We think about three audiences: the coding agent, the developer, and then the engineering leader. And the way we think about it is that if everything’s running perfectly, if code is flowing from plan to main at the speed of agentic development, the agents are in service of the developers, and then the developers are in service of the engineering leader.” He said it can be thought of as the agents are the frontline capacity of production, whether it’s for planning, or coding, or review, but ultimately they “report” to the developers who are orchestraing that work, signing off on it, or intervening on the work when necessary. Then engineer reports to a manager, who reports into a VP, who wants code flow because they want to run 50% faster. “It’s all the things around code generation that create friction to go from, say, Claude Code generated some code, so what are all the things that need to be true for me to ship this into production in a safe and trusting way?”
The framework focuses on the critical elements of visibility, governance, and integration. Because agents are creating a massive explosion in the speed and volume of code, teams need a better way to understand where work is getting stuck and how individual agents are performing. Code Flow is the discipline of maintaining standards and keeping context clear as AI-driven development becomes the standard.
Introducing Kepler
GitKraken recently launched new products through which code can flow. The first is Kepler, an purpose-built agent development environment (ADE), which lives on the desktop for Windows, Mac and Linux. The company also launched GitKraken Desktop 12 with a new agent mode, and GitLens 18 — a popular IDE extension in the VS Code Cursor and anti-gravity ecosystem — which adds agentic capabilities.
Johnston said Kepler is part of what the company calls its developer experience with Kepler, GitKraken Desktop and GitLens. “GitKraken Desktop and GitLens were built to manage Git and help unwind conflict resolutions, and have greater code visibility,” he said. “They weren’t built for, ‘How do I deploy 20 agents at once, who turn around and spin up 200 sub agents?’
From plan to main
All the system around code production weren’t build for today’s volume or velocity, and includes everything from code review to conflict resolution to code context, to generating the right type and size commits with the right level of description. Johnston went on to describe how customers would use the Code Flow framework. “You can work in kind of traditional Git mode, that’s GitKraken Desktop. If you’re working in code mode, that’s in your IDE, that’s GitLens. And if you are agent-first mode, that’s Kepler. And what we find early on from our preview users that have been using Kepler for the last few weeks is it’s not one or the other. They’ll often say, ‘I actually start my day or start my work in Kepler, I spin up my 12 agents, I’m doing x, y, and z, and then I might realize that I need to go refactor some code. I go through GitLens, through Cursor into my IDE, and then before I can merge, I realized I’ve got a big kind of a nasty conflict. Then I’m opening GitKraken Desktop, which handles things like conflict resolution really well. That’s all part of the reason that we include all of them in one plan and one subscription, because they are more modes of work for a developer than they are three separate products.”
The post GitKraken Unveils Code Flow to Help Teams Navigate the AI Era appeared first on SD Times.
Tech Developers
No comments