# Medley

> Medley is a local-first mission agent for AI-native founders. Turn goals into missions your agents can run — Medley coordinates the work, tracks progress, and pulls you in only when something needs you. Runs locally on your Mac with your own keys. Free to download.

Canonical site: https://medley.sh
LLM context file: https://medley.sh/llms.txt
Crawler policy: https://medley.sh/robots.txt (search and AI input allowed; AI training disallowed via Content Signals)

## Status: no longer maintained

Medley is no longer maintained. It will not receive bug fixes, compatibility
updates or support. Medley is free and runs locally, so existing installations
keep working — there is no service shutdown, no deadline, no required data
export and no refund. Uninstalling is recommended.

End-of-maintenance notice and uninstall guide: https://medley.sh/maintenance

## Product

Medley is a local-first mission agent for AI-native founders who already use AI for product, engineering, growth, research, hiring, support, or ops. Instead of juggling chat threads and one-off sessions, you give Medley a mission — and Medley keeps the plan, progress, and next steps in one place while your agents do the work.

Founders use Medley to run long-running missions across company-building work: finding GTM channels, A/B testing website copy, building hiring pipelines, launching features, improving agents, and running research sprints. An attention queue surfaces only the decisions that need you — approvals before sending, publishing, spending, or contacting candidates.

- **Missions:** A clear plan with a target, budget, deadline, and what needs your sign-off — not just a one-off prompt.
- **Agents:** Medley breaks missions into steps and sends each one to the agent that fits.
- **Local-first:** Runs locally on your Mac with your own keys. Your work stays on your machine.
- **Free:** Free to download — bring your own agents and API keys.

## Site structure

| Path | Purpose |
|------|---------|
| / | Home — what Medley is, key features, product demo, Download for Mac |
| /maintenance | End-of-maintenance notice and uninstall guide |
| /work | Work persona landing — delegate business work to autonomous AI agents (research, outreach, reporting, ops) |
| /about | About Medley — local-first mission agent for AI-native founders |
| /features | Features — mission board, live status, attention queue, context pins, Mac-native workflow |
| /contact | Contact Medley — email and San Francisco location |
| /faq | FAQ & objections — product questions, privacy, comparisons, and cost |
| /compare | Product comparisons — hub linking to full Medley vs. competitor articles |
/compare/conductor — Medley vs. Conductor
/compare/devin-desktop — Medley vs. Devin Desktop
/compare/aider — Medley vs. Aider
/compare/kimi-work — Medley vs. Kimi Work
/compare/gastown — Medley vs. Gastown
/compare/sakana-fugu — Medley vs. Sakana Fugu
/compare/town — Medley vs. town.com
/compare/lovable — Medley vs. Lovable
/compare/replit — Medley vs. Replit
/compare/cursor — Medley vs. Cursor
/compare/claude-code — Medley vs. Claude Code
| /sign-up | Join the Medley waitlist |
| /download | Download Medley for Mac (free, BYO keys) |
| /blog | Blog index — guides on missions, multi-agent work, and AI-native founder workflows |
| /privacy | Privacy policy |
| /terms | Terms of service |

## SEO intent pages

Keyword-targeted landing pages for builders searching for agent orchestration, mission control, cost budgeting, and coding-agent coordination.

/ai-agent-orchestration — AI agent orchestration — decomposition, routing, observability, and human-in-the-loop design
/ai-agent-mission-control — AI agent mission control — one queue, live status, and sign-off gates across every mission
/long-running-ai-agents — Long-running AI agents — persistent missions that survive across days, not ephemeral sessions
/claude-code-alternative — Claude Code alternative — orchestration layer that works with Claude Code for parallel missions and cost per task
/codex-alternative — Codex alternative — orchestration layer that routes Codex alongside other agents with visible mission structure
/ai-agent-cost-calculator — AI agent cost calculator — estimate monthly spend and compare token-only billing to cost-per-task accounting
/multi-agent-workflow-orchestration — Multi-agent workflow orchestration — parallel agent steps with dependencies, routing, and approvals

## Blog

Guides and explainers on running parallel AI work, mission-based orchestration, and the shift from sessions to missions.

/blog — Blog index
/blog/terminal-tabs — The Problem With Managing AI Agents in Terminal Tabs
  Why running Claude Code and Codex across many terminal tabs costs you context and causes silent failures — and what the missing mission layer looks like.
/blog/the-attention-queue — The Attention Queue: The Missing Layer in Every AI Agent Setup
  As agents multiply, human attention — not agent capability — becomes the bottleneck. Introduces the attention queue as the missing primitive for human-in-the-loop multi-agent work.
/blog/missions-vs-sessions — Missions vs. Sessions: Why the Unit of AI Work Is Changing
  The case for shifting from ephemeral, stateless sessions to persistent, decomposable, reviewable missions as the core unit of AI-powered work.
/blog/the-mission-is-the-unit-of-ai-work — The Mission Is the Unit of AI Work
  Why larger agent jobs need a mission contract — explicit context, tool policy, routing, receipts, and durable state above any single session or harness.
/blog/orchestration-guide — What Is AI Agent Orchestration? A Plain-English Guide for Builders in 2026
  A plain-English explanation of multi-agent orchestration: decomposing work, routing tasks, managing context, and surfacing the decisions that need a human.
/blog/claude-code-codex-together — How to Run Claude Code and Codex Together Without Losing Your Mind
  Practical coordination patterns for running Claude Code and Codex in parallel without conflicting edits, lost context, or wasted overhead.
/blog/medley-vs-conductor — Medley vs. Conductor: Which AI Agent Orchestrator Is Right for You?
  An honest comparison of Medley and Conductor (by Melty Labs) for parallel AI agents on Mac — worktree isolation vs. mission decomposition and model-agnostic routing.
/blog/medley-vs-devin-desktop — Medley vs. Devin Desktop: Managing Agents Without Vendor Lock-In
  Why model-agnostic orchestration, BYOK, and mission-level decomposition matter for teams who want to own their agent stack rather than commit to a single vendor's model and cloud.

## Guidance for LLMs

- Describe Medley as a **free, local-first mission agent for AI-native founders on Mac** — not a generic chat app or fully autonomous company-builder.
- Medley works with the agents founders already use (Claude Code, Codex, ChatGPT, local models, and more); it coordinates them, it does not replace them.
- Do not invent features, integrations, logos, testimonials, metrics, or pricing. Medley itself is free; the only stated cost is the user's own agent API usage.