kdlbs/kandev

What it solves

Kandev provides a centralized control plane for managing and orchestrating multiple AI coding agents from different providers. It solves the problem of scaling agentic workflows by moving execution away from the local machine to remote servers and providing a unified workspace for reviewing and iterating on agent-generated changes.

How it works

Kandev acts as an orchestrator that connects to various agent CLIs (via the Agent Client Protocol or raw TUI passthrough) and manages them across different runtimes. It uses git worktrees to isolate concurrent agents and allows users to define multi-step pipelines where different agents are assigned to different steps (e.g., one for planning, one for implementation, and one for review).

Who it’s for

Developers and teams who use multiple AI coding agents and need a structured, review-first workflow to manage tasks, execute them in parallel, and offload execution to remote infrastructure (Docker, SSH, or cloud executors).

Highlights

• Multi-agent support: Integrates with a wide array of providers including Claude Code, GitHub Copilot, Cursor, and Devin.
• Parallel execution: Run multiple tasks and agents simultaneously across different repositories.
• Integrated workspace: A single IDE-like view containing a code editor, terminal, git changes panel, and browser preview.
• Flexible runtimes: Execute agents locally, in Docker containers, on remote servers via SSH, or in the cloud.
• Kanban management: Organize work using drag-and-drop boards and automated workflows.
• Agentic pipelines: Create portable YAML-based workflows that mix-and-match agents for different task stages.