The short answer
Pick the lightest tool that solves your problem.
If you run one agent and only need to know when it stops, a sound hook or terminal-notifier is enough, and every page below says so. Heard is for the next problem: several Claude Code or Codex sessions, and wanting to hear what changed and what needs you without going back to the screen.
All comparisons
Pick the question closest to yours
macOS coding-agent notification tools comparedFourteen ways to follow Claude Code and Codex, from simple pings to spoken updates you can answer.↗
Heard vs. bells, banners and generic text-to-speechWhen a simple notification is enough, and when context matters.↗
terminal-notifier for Claude Code and Codex.Set up terminal-notifier for Claude Code and Codex, put the project name on every banner, fix banners that never appear, and add speech if you want it.↗
How to make Claude Code talk: every text-to-speech option compared.Claude Code has voice input but no spoken output. Compare a say hook, claude-code-narrator, claude-code-tts, VoiceMode, Aloud and Heard, with setup for each.↗
Spoken notifications for OpenAI Codex.Hear when OpenAI Codex finishes or needs approval: Heard speaks it on macOS, or use the notify command with say. Also agent-notify, Voxlert, Agent Voice Reply.↗
Heard vs. AgentNotch: a board you glance at, or a voice you hear.AgentNotch shows your coding agents in the Mac notch; Heard tells you about them out loud. How they differ, plus AgentNotch alternatives by job.↗
Open-source voice tools for coding agents, and what really runs locally.Ten open-source voice tools for Claude Code and Codex compared, from Heard and OpenLive to VoiceMode and dictare, and which ones send text to the cloud.↗
What a voice layer for Claude Code is, and the two ways to build one.What is a voice layer for Claude Code? Compare Claude Code’s /voice, VoiceLayer’s MCP tools, VoiceMode and Heard’s hook-based voice layer.↗
Tools that read coding agent progress aloud, compared.Heard, Aloud, Crier, Agent Hotline, Terminal Talk, AgentVibes, Agents Aloud, AgentsRoom and Orb compared: what each reads aloud, where the voice runs, platforms, price.↗
Heard vs. VoiceMode: talk with one agent, or hear from all of them.VoiceMode lets you talk with Claude Code; Heard speaks your agents’ progress from their hooks. How they differ, where each wins, and setup.↗
Heard vs. Agents Aloud: hear the agent’s own words, or a short update.Agents Aloud reads your Claude Code and Codex messages aloud in a Mac app; Heard speaks short updates from their hooks. How they differ and when to pick each.↗
Heard vs. Aloud: per-session alerts, or one voice for every session.Aloud and Heard both speak Claude Code and Codex alerts aloud on macOS. How they differ in what they read, local voices, setup, price, and when to pick each.↗
Heard vs. AgentVibes: the agent speaks for itself, or a voice layer speaks for it.AgentVibes has your agent speak when a task starts and ends; Heard speaks for the agent from its hooks. How they differ in words, voices, platforms and price.↗
Heard vs. Crier: two ways to hear Claude Code and Codex progress aloud.Crier and Heard both speak up when Claude Code or Codex needs you on a Mac. What each says, local voices, agents, setup, price, and when to pick each.↗


