# ManyAgents ManyAgents is a native Mac app for coordinating Claude Code and Codex agents across projects, with separate task tabs and live web or iOS previews beside the chat. - Findings24: - Website: - Pricing: Freemium - Categories: AI Agent Infrastructure ## What is ManyAgents? ManyAgents is a native Mac app for coordinating multiple Claude Code and Codex coding agents from one window. Assign a project request to an orchestrator, which breaks it into tasks and tracks worker agents in separate tabs. Agents use the user's existing Claude Code and Codex access, and project code remains on the Mac. A built-in preview panel supports checking web projects and viewing iOS apps in the iOS Simulator. ## What can ManyAgents do? ### Project orchestrator Give an orchestrator a project request; it splits the work into tasks, starts agent tabs and tracks who is handling each task. ### Isolated worker worktrees Workers build assigned tasks in separate Git worktrees, then test their work and report back. ### Specialist thinkers and reviewers Thinkers can be called for difficult decisions or cross-codebase problems. Reviewers check completed work with read-only access; when Codex is configured, reviewers can run on a different provider. ### Project and agent card view A single canvas shows projects, their orchestrators, agent assignments and worktrees, with status indicators for active, queued or waiting work. ### In-app web testing Agents can open the app in a browser panel, interact with pages and forms, and inspect results. The preview can stay signed in, and users can take over with a click. ### iOS Simulator panel Switch the preview panel to the iOS Simulator to view an iOS app alongside the agent working on it. ## Use Cases ### Split a project request into parallel tasks Have an orchestrator assign distinct implementation tasks to workers, each operating in its own worktree, while keeping progress visible in the project view. ### Check a web checkout flow Ask agents to exercise a web app workflow such as creating a new account, adding an item to a cart and submitting an order, then inspect the rendered result. ### Review a difficult code decision Bring in a thinker for a problem spanning the codebase or one that workers have not resolved, and route the resulting decision back to workers. ### Inspect an iOS build during development Use the in-window iOS Simulator panel to view an app screen beside the coding conversation rather than building to a physical phone for that check. ## Frequently Asked Questions ### What do I need to run ManyAgents? A Mac running macOS 14 or later and Claude Code with a Claude Pro or Max subscription, or an Anthropic API key. Codex is optional. ### Does ManyAgents provide its own AI access? No. Agents run through the user's own Claude Code and Codex access; the site says there is no extra AI bill from ManyAgents. ### Where does project code run? The product states that agents run on your Mac and your code is not sent to ManyAgents. It also says the app has no analytics. ### What happens when the 14-day trial ends? You can continue to read existing work, but new work pauses until you add a license. The site lists a one-time $49 license, purchased by emailing hello@ailogy.co. ### Can agents test my web app? Yes. Agents can use the built-in browser preview to click through pages, fill forms and check results; you can take over the preview with a click. ## Quick Facts - **Product type**: Native Mac app for coding-agent coordination - **Platform**: macOS 14 or later - **Supported coding agents**: Claude Code and optional Codex - **Trial**: 14 days - **Code location**: Runs on the user's Mac - **Website**: manyagents.app