Medley is a Claude Code plugin that transforms a single prompt into a coordinated multi‑agent plan, generating a URL to visualize the workflow. By orchestrating agents sequentially, in parallel, or overnight, it handles complex tasks such as research, outreach, documentation, and integration with tools like GitHub, Slack, Notion, and AWS, delivering results that surpass what a single model can achieve.
Funding
Funding not disclosed
Founders
Product
Problem
Coordinating multiple AI agents across tools, sessions, and long‑running tasks requires users to manually track context, approvals, and tool permissions, leading to fragmented workflows and hidden costs. This makes it difficult to execute complex projects such as research, documentation, or outreach that extend beyond a single prompt.
Solution
Medley is a Claude Code plugin that transforms a single user prompt into a structured, multi‑agent mission. It automatically decomposes the request into a visible plan, assigns tasks to appropriate agents, and runs them sequentially, in parallel, or overnight. The plugin integrates with a wide range of development and productivity tools—including GitHub, Slack, Notion, AWS, and many others—so agents can act directly within the user’s existing ecosystem. Users receive a URL that visualizes the mission’s structure and progress, while judgment gates allow human approval before critical actions such as publishing or spending. By providing persistent context (pinned files, notes, and decisions) and a unified status view, Medley keeps long‑horizon AI work coordinated without requiring a separate orchestration app.
Target Audience
Medley is aimed at developers, product teams, and operations professionals who use Claude Code and need to automate complex, multi‑step AI‑driven workflows across integrated tools.
Features
- Automatic prompt decomposition into a detailed, multi‑agent execution plan with a shareable URL for real‑time monitoring
- Support for sequential, parallel, and overnight agent execution to handle tasks of varying duration
- Native integration with over 30 popular tools (e.g., GitHub, Slack, Notion, AWS, Google Workspace, Docker, Jira) enabling agents to act directly in existing workflows
- Persistent mission context that pins files, notes, and decisions, preventing agents from starting cold on each step
- Human‑in‑the‑loop approval gates that pause execution before publishing, sending communications, or incurring costs
- Mac‑native status dashboard that aggregates progress, logs, and agent outputs in a single view within Claude Code
- No separate application required; all orchestration occurs inside the Claude Code environment via the /mission command