Kadmo

Kadmo is an AI coding agent platform for agentic engineering: it runs Claude Code, Codex and your other AI coding agents as one team on your own cloud. Each agent works in its role, on its own machine and inside your tools, and nothing moves on before the gate at its handoff passes. You drop a ticket and the fleet takes it from there: a playbook carries the ticket through its steps, the product manager agent splits the work into tasks sized for one agent, the software engineer agent opens the pull request and fixes what review sends back, a senior reviews, and the QA agent tests the result on the live site. Every step writes into the ticket what it did and why. Roles: Software Engineer, Product Manager, QA Engineer, Documentation, Skill Discovery and Website Builder. Tools: Jira, Linear, Notion, GitHub, GitLab, Slack and Microsoft Teams. Cloud: AWS, Google Cloud or Hetzner. Models: Claude, GPT and Mistral, or on Enterprise your own keys and gateway.

Kadmo coordinates a fleet of role-based agents to work on engineering tasks in a team's existing tools and cloud environment. Agents take tickets, follow playbooks, and hand work between roles such as product management, software engineering, and QA. The platform is designed to return results in the ticket, including pull requests, test outcomes, and explanations of work performed. It addresses the need to delegate engineering work while keeping review, approvals, and operational controls in place. A team connects a task board and repository, then gives the fleet a ticket or goal. Playbooks divide work into steps for agents in defined roles, with gates and approvals at handoffs. Agents work on their own machines and branches, open pull requests, and address changes requested in review. QA can run a test suite on the live environment and file failures as tickets. Each step leaves a comment explaining what the agent did and why, and runs retain a replayable transcript. Kadmo is for engineering teams that want agents to work within their existing task boards, repositories, cloud environments, and model setup. The page describes teams using roles such as product manager, software engineer, and QA engineer, with additional agents for documentation, skill discovery, and website building. It emphasizes scoped credentials, configurable approval gates, budget limits, data residency policies, and audit records, while allowing teams to use their own cloud or machines.