kandev: A powerful tool for deeper control over how AI agents work through customizable workflows, agent profiles, runtimes, prompts, and review gates.
What problem it solves
AI agents are often run in terminal interfaces, but reviewing and iterating on their changes there does not scale. Teams also need a way to organize parallel work, coordinate agents, and keep human review at the center of the development process.
Key features
Parallel task orchestration — Run and manage multiple tasks from different sources at the same time, including tasks that span multiple repositories and produce separate pull requests.
Source: README
Integrated review workspace — Review agent output through a workspace that combines a file editor, file tree, terminal, browser preview, git changes, and chat in one interface.
Source: README
Configurable agentic workflows — Define reusable workflows and pipelines, assign different agents to individual steps, and apply review gates that fit the team's process.
Source: README
Flexible execution environments — Run agents as local processes, in isolated containers, on remote servers over SSH, or in cloud executors, with reusable profiles for runtime configuration.
Source: README
Self-hosted, multi-provider platform — Use supported agent CLIs from different providers, run the application locally or on your own infrastructure, and access it remotely without telemetry.
Source: README
Quick start
Copied verbatim from the project's README.
brew install kdlbs/kandev/kandev
kandevnpx kandev@latestnpm install -g kandev@latest
kandevWho it may suit
Inference: this may suit developers and teams that run several AI coding agents, need structured review gates, or want to execute agent workloads across local and remote infrastructure.
Topics to discuss
Which workflow and review-gate defaults would help teams adopt a shared agentic process?
Which executor and task-isolation features are most valuable for large codebases?
What integrations or workflow extensions should be prioritized next?
Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.
Read the original README https://github.com/kdlbs/kandev#readme
ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.
Comments
…