Back to Gap Map Product / UX / Orchestration & agents

agent-manager

open source / Overall score: 1.7

Terminal UI that runs coding-agent CLIs side by side, each in its own persistent tmux session on a private tmux server. It launches the user's own installed CLI unmodified, shows every session's live status in one foldable project tree, answers a session without attaching to it, revives or forks a session on its own conversation, and opens a full-file diff review whose line comments go back to the agent as one prompt. It supports Claude Code, Codex, OpenCode, Grok Build, Gemini CLI, Pi, Command Code, Hermes Agent and Muse Code, and ships as a Go binary for macOS, Linux and Windows under WSL2.

Openness

5 high confidence

5.0

license: Apache-2.0(OSI)
source: public
core features withheld: no

Everything agent-manager does is in the public repository, and nothing is sold beside it. The coding CLIs it launches bill through their own vendors.

https://agent-manager.dev/pricing.md recorded 2026-09-30
one tier, free under Apache-2.0: the terminal UI, its built-in MCP server and the reference API all free with no account, key or quota; only the coding CLIs it launches bill, through their own vendors
https://raw.githubusercontent.com/YoanWai/agent-manager/main/LICENSE recorded 2026-09-30
Apache License 2.0, the standard license body with the appendix boilerplate unmodified
https://raw.githubusercontent.com/YoanWai/agent-manager/main/README.md recorded 2026-09-30
README with the full Go source building from the repository (go run ., go test ./...) and an Apache-2.0 license section

Adoption

1 low confidence

1.0

GitHub stars are the signal read for agent-manager, which ships as a Go binary through Homebrew, an install script and release downloads rather than through one package registry. A star is not a use, so this shows attention rather than usage.

https://img.shields.io/github/stars/YoanWai/agent-manager.json recorded 2026-09-30
shields.io relay of the GitHub stargazer count for YoanWai/agent-manager: 543 stars

Capability

2 medium confidence

2.0

agent-manager supervises coding agents the user already runs rather than being one, and sits with YYLO, the other layer of this kind in the category: YYLO structures the task pipeline, agent-manager the live sessions and their review. It is capable within that scope, and has no third-party benchmark coverage.

https://raw.githubusercontent.com/YoanWai/agent-manager/main/README.md recorded 2026-09-30
README: live status for nine supported coding CLIs in tmux, prompts sent without attaching, revive and fork, full-file diff review with comments returned to the agent, per-session worktrees, and MCP tools for agents to spawn and message each other

Verified 2026-09-30