Conductor by Melty Labs
Medley vs. Conductor
Conductor is an excellent Mac-native tool for parallel coding agents. Medley is built for founders who need missions that span coding and adjacent company-building work — with decomposition, routing, and one attention queue.
Paste into Claude Code
If you run multiple AI coding agents on a Mac, Conductor and Medley are both worth evaluating — and the honest answer is that they solve different problems.
Conductor’s insight is powerful: parallel agents need isolated workspaces. Medley’s insight is different: parallel work needs a project layer — decomposition, routing, and a queue for what actually needs you.
What Conductor does well
Conductor spins up separate Git worktrees for each agent session so agents do not step on each other’s changes. The diff, review, and merge UX is polished — you compare outputs and merge the best result without leaving the app.
For a developer running pure coding tasks inside one codebase — write this feature, fix this bug, refactor this module — Conductor delivers a clean, Mac-native experience with low setup friction and the models you already have passthrough for Claude Code keys.
Where Conductor stops
Conductor is built around the assumption that you are the planner. You decide what to run, which agents to assign, and how to sequence work. It executes in parallel — it does not decompose missions across domains.
There is no cross-project attention queue, no cost per finished task, and no decision memory that compounds your approvals over time. If your work spans code plus a changelog, a GTM doc, and outreach — or multiple projects at once — you are still the integration layer.
What Medley does differently
Medley starts from the mission, not the task list. You describe the outcome; Medley is designed to produce a visible, editable DAG of sub-tasks and route each step to the right agent and model tier.
An attention queue is designed to surface only the decisions that need you across every project. Cost is meant to attach to finished tasks, not a monthly token blob. Missions can span coding and adjacent company-building work — the project layer Conductor does not provide.
Key differences
Mission decomposition vs. manual task lists
Medley decomposes goals into a DAG. Conductor parallelizes the list you build yourself.
Cross-domain missions
Medley runs coding and adjacent company-building work in one mission. Conductor is code-only.
Attention queue
Medley pulls you in only when needed. Conductor relies on you checking each session’s diff dashboard.
Cost per task
Medley attributes spend to finished work. Conductor reports tokens through your agents, not per outcome.
Full product comparison
| Medley | Conductor | |
|---|---|---|
| Product type | Project orchestrator | Desktop multi-agent orchestrator |
| Coding vs. non-coding work | Both — coding and adjacent company-building work | Coding only |
| Project layer | Project-level — task decomposition across the project | Session-level — parallel sessions, not project-wide |
| Attention queue / human-in-the-loop | Centralized queue — humans pulled in only when needed | Visual diff dashboard — no queue, manual review |
| Model access | Uses models you already have in Claude Code | Yes — passthrough for Claude/Codex keys |
| Local vs. cloud | Local — Claude Code plugin | Local-first — macOS desktop |
| Model support | Models you bring into Claude Code (Claude, GPT, Gemini, and more) | Claude Code, Codex, OpenCode (multi-model mode) |
| Platform / OS | Mac — desktop app | macOS only — Apple Silicon |
| Pricing & business model | Free — models you already have in Claude Code | Free app — BYOK, pay own API costs |
| Target user / ICP | Founders & power users of Claude Code | Developers running multiple AI tasks in parallel |
Data as of June 2026. Sources: web research, medley.sh.
Which should you choose?
Choose Conductor if
- Your work is primarily or exclusively coding inside one codebase
- You want worktree isolation and polished diff/merge UX
- You have a clear task list and want to parallelize execution
- You prefer a minimal, focused tool with low setup overhead
Choose Medley if
- Your missions span code, content, research, and coordination
- You run multiple projects and need one attention queue
- You want the system to decompose work and route across models
- You need cost visibility per finished task, not just per token
Medley vs. Conductor FAQ
- Is Medley a replacement for Conductor?
- Not directly. Conductor is a coding orchestrator with worktree isolation. Medley is a project orchestrator that coordinates Claude Code and other agents across full missions. Some builders use concepts from both categories depending on workflow.
- Can I use both tools?
- They overlap on parallel coding but serve different layers. Conductor excels at isolated coding sessions; Medley excels at multi-domain missions and cross-project visibility. Most teams pick the layer that matches how they actually work.
- Are both Mac-native?
- Yes. Both target Mac power users. Conductor is macOS-only on Apple Silicon. Medley is a local-first Mac app in Claude Code — install at medley.sh/install.
Outgrown the task-list model?
Install Medley — mission orchestration for founders who run Claude Code across real company-building work.
Install MedleyI