Kepler

Product Introduction

1. Definition: Kepler is GitKraken's Agentic Development Environment (ADE), a specialized software platform designed to orchestrate, manage, and review the output of multiple AI coding agents working in parallel across one or more code repositories.
2. Core Value Proposition: Kepler exists to solve the emerging developer workflow challenge of "agent sprawl," providing a unified interface for developers to scale their use of AI agents for code generation, review, and implementation, thereby maximizing developer productivity and maintaining code quality oversight.

Main Features

1. Task-Based Agent Orchestration: Kepler introduces the concept of a "Task" as a top-level container for work that can span multiple repositories. It integrates directly with issue trackers (e.g., Jira, Linear, GitHub Issues) and pull request queues. A developer can select an issue or PR, and Kepler automatically sets up the necessary environment, creates worktrees, and launches pre-configured AI agents (such as Claude Code, Codex, or Open Code) to execute the work. This "start from your backlog" functionality eliminates manual context-switching and session setup, connecting your existing workflow directly to agent execution.
2. Kanban-Style Visibility Dashboard: The core of Kepler's management interface is a Kanban board that tracks every agent session and Task from "Exploration" to "Done." Each card represents an active agent session, displaying critical metadata: the assigned branch, parent Task, target repository, and last agent activity. Sessions are filtered by status (Needs Attention, Active, Idle, Errored, Inactive), enabling developers to instantly assess the pipeline's health and focus their attention where intervention is required.
3. Integrated Agent Console & Control: For each running agent session, Kepler provides a dedicated "Console" panel. This is a real-time log of the agent's actions, thoughts, and code changes. Crucially, it includes an input field for direct communication, allowing a developer to type instructions, paste screenshots, or even use voice commands to course-correct an agent mid-task. Multiple consoles can be opened side-by-side for comparative analysis across concurrent agent activities in different repositories.
4. In-Context Code Review and Commit: Kepler seamlessly bridges the gap between agent execution and code integration. Within the Task or session view, a developer can directly open a worktree to see the Git diff of all changes made by the agent. The interface supports staging files, writing commit messages, and finalizing commits—all without leaving the Kepler environment. This creates a closed-loop workflow from agent initiation to committed code.
5. Agent-Agnostic & Ecosystem Integration: Kepler is designed as a "surface" layer that sits atop existing tools. It is agent-agnostic, supporting integration with popular AI CLI tools like Claude Code and Codex. It connects to major Git providers, including self-hosted services like GitHub Enterprise and GitLab Self-Managed (via Personal Access Tokens). The platform also integrates with the broader GitKraken suite (Desktop, CLI, GitLens) and engineering management tools (e.g., GitKraken Insights), positioning it within a complete developer experience (DevEx) platform.

Problems Solved

1. Pain Point: Agent Sprawl and Loss of Visibility. As developers use more AI agents to accelerate coding, they face a chaotic workflow involving multiple terminals, separate contexts, and no unified way to track what each agent is doing, what changes have been made, or which agents require attention. This leads to duplicated work, missed errors, and inefficient oversight.
2. Target Audience: The primary users are AI-augmented developers, engineering team leads, and AI architects who are responsible for integrating and scaling AI coding assistants within their team's workflow. This includes roles like Lead AI Engineers and Senior Software Architects who need to manage agent deployment across complex, multi-repository projects.
3. Use Cases: Kepler is essential for scenarios including: orchestrating a multi-agent workflow for a feature that requires simultaneous changes to a database schema, API, and front-end; managing dozens of concurrent AI agents performing code refactoring across a microservices architecture; providing a centralized review and quality gate for all code generated by AI agents before it enters the main codebase; and streamlining the handoff of tasks from project management tools (like a Jira ticket) to automated coding agents.

Unique Advantages

1. Differentiation vs. Traditional IDEs and Generic AI Tools: Traditional Integrated Development Environments (IDEs) are built for a single developer writing code line-by-line. Generic AI coding assistants often operate as single agents within a single repo context. Kepler differentiates itself as an Agentic Development Environment built explicitly for the paradigm of one developer directing many AI agents across multiple repositories. It replaces fragmented terminal sessions and manual orchestration with a single, managed dashboard for parallel agentic work.
2. Key Innovation: The ADE Paradigm and Context-Linked Workflow. Kepler's key innovation is formalizing the "ADE" category, shifting the developer's role from code writer to agent orchestrator. Its most significant technical innovation is the deep linking between the source of work (a Jira issue or PR), the generated Task, the agent sessions spawned to execute it, and the resulting code changes. This context continuity is maintained across the entire workflow, eliminating the "context rebuild" problem common when moving work between different AI and development tools.

Frequently Asked Questions (FAQ)

1. What is the difference between Kepler and a traditional IDE or a tool like GitHub Copilot? Kepler is not a replacement for an IDE or a single-agent coding assistant. It is an orchestration and management layer built to scale AI coding agents. While an IDE is for writing code and Copilot might suggest a single line, Kepler is for directing multiple autonomous agents to complete entire features across several repositories, while providing the dashboards to manage and review their collective output.
2. Which AI coding agents does Kepler support? Kepler is agent-agnostic and currently integrates with popular AI CLI workflows, including Claude Code, Codex, and Open Code. Its design allows you to connect the agents your team is already using, providing a consistent management interface regardless of the specific underlying model or ecosystem.
3. How do I start using Kepler? Is it free? Kepler ADE is available as a free download for Windows, Mac, and Linux. During its limited preview period, it is free for all users. For long-term access and to secure additional features, you can explore the GitKraken Pro, Advanced, or Business plans. You can start by downloading Kepler, connecting it to your Git provider and issue tracker, and launching your first Task from an existing issue or pull request.
4. Can I use Kepler with my self-hosted GitHub Enterprise or GitLab instance? Yes, Kepler supports GitHub Enterprise and GitLab Self-Managed. Integration requires generating a Personal Access Token (PAT) with the appropriate permissions for your repositories and configuring it within Kepler's integration settings.
5. How does Kepler handle security and code control? Kepler provides centralized control and visibility over all agent-generated code. Developers must review agent changes via the integrated diff tool before committing. The platform is designed for enterprise-grade needs, and the broader GitKraken suite includes advanced security controls and compliance features, making it suitable for engineering leaders managing team velocity and risk.