Lody

What is Lody

Lody is a shared workspace for coordinating AI coding agents across a team. It is built around isolated Git worktrees, so separate agent sessions can work in parallel without stepping on each other’s files. The product also emphasizes live diff review, which lets you inspect changes as they happen instead of waiting for a finished handoff.

The app is designed as a control plane for coding automation rather than a single-agent chat. You can open sessions, share them with teammates, and keep the same context while moving between desktop, browser, and mobile. Lody’s own material describes support for GitHub workflows, PR status, checks, conversation, and merge actions inside the session flow.

Key features

• Parallel worktrees for running multiple AI coding tasks at the same time.
• Shared sessions so teammates can open the same thread and continue the conversation.
• Live diff review for watching code changes as they land.
• GitHub integration for pull request status, checks, comments, and merge-related actions.
• Mobile access, including an on-the-go experience with Dynamic Island progress and permission prompts.
• Support for many agent tools through the Agent Client Protocol, including Claude Code, Codex, Cursor, Gemini CLI, Cline, Goose, OpenCode, Qwen, and more.
• API key support, which means you can connect agents that use keys rather than only interactive logins.
• Usage tracking by member, with token and cost breakdowns by model and teammate for the workspace.
• Multi-platform sub-agents for testing across macOS, Linux, and Windows in parallel.
• Cross-repo workflows where one session can help fix a dependency in another and then bring the result back.

How to use it

1. Start by opening the web app or downloading the desktop option if you want a local workflow.
2. Connect a machine by running npx lody daemon start on a server, cloud VM, or desktop that should stay online.
3. Sign in through the browser link the daemon opens, then keep that machine attached for future sessions.
4. Create a session from the terminal, for example with lody session create --agent codex "Fix the auth 500s".
5. Chat with an active session using lody session chat <id> and keep steering the agent as work progresses.
6. Use shared sessions to let teammates review diffs, check status, and continue the same thread.
7. Add GitHub context when you are working on pull requests, CI failures, or merge steps.

Who it is for

Lody is aimed at engineering teams that already use coding agents and want a more structured way to run them. It is especially relevant for:

• product teams that need several agent tasks moving at once,
• developers working across multiple repos or dependencies,
• teams that want a review-then-fix loop without copying context between tools,
• engineers who want to keep sessions visible to collaborators,
• people who need to monitor agents from phone, browser, and desktop.

It also fits teams that prefer bringing their own agent subscriptions or login-based tools, since Lody says it can run on subscriptions you already have.

What to know before you use it

Lody is not a generic code editor, and it is not described as a standalone model provider. It acts as the orchestration layer around external coding agents, so the quality of results still depends on the agent you connect, your repo setup, and the permissions you grant.

A few practical caveats are worth noting:

• The site says machines stay private until you share them, so session visibility is a workspace choice rather than automatic public collaboration.
• Usage and cost reporting are tied to the workspace and the connected models, so you should review those numbers if you are tracking spend.
• GitHub actions appear to be part of the workflow, but the public material does not specify every supported hosting provider or enterprise setup.
• The app highlights support for ACP-compatible agents, but exact behavior can vary by tool and by how well that tool adapts to the Agent Client Protocol.
• For final source-of-truth work such as merge decisions, pull request checks, or repo history, use your official GitHub records and the connected machine state rather than relying on a chat summary alone.

If you need a team workspace for managing AI coding agents across devices and repositories, Lody is built for that kind of coordinated development work.