Heard is live · pricing from $24/mo · last verified 8 Nov 2026 · viability healthy
Heard
Heard is a voice context layer that narrates your Claude Code and Codex agent sessions aloud so you can stop watching the terminal.
If your Claude Code or Codex runs are long and you keep flicking back to the terminal, Heard converts that reflex into audio you can act on. Power is the tier that changes the workflow, since Parrot conversations and iPhone pickup are where hands-off actually happens. The 14-day card-free Power trial makes the paid step cheap to test.
Verdict generatedVerified 5d ago · liveness 76/100 · cite: rightaichoice.com/tools/heard
- Mac developers running long Claude Code or Codex sessions who keep tab-switching to check progress
- Engineers who step away from the desk and want iPhone text, voice-note, or call pickup on Power
- Builders who want an open-source, inspectable voice layer with local Kokoro narration
- Teams needing activity across projects, admin controls, and zero data retention on the Team tier
- Windows or Android users: Heard supports macOS 14 or later, with no Windows/Android today
- Anyone whose agents fall outside the supported list and won't use the open source heard run wrapper
- Silent shared offices where speaking to a hotkey isn't practical
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Heard if you are on Windows or Android, your agents are outside the Claude Code, Codex, or MCP cloud-agent list and you won't build the open source engine, or your agent runs finish in seconds so narration overhead never pays for itself.
Our "biggest gripe" note written Oct 15, 2026 quotes $12, $144, which no longer matches the verified price table, so it is being re-reviewed and is not shown.
Our price-reality note written Oct 15, 2026 quotes $12, $144, which no longer matches the verified price table, so it is being re-reviewed and is not shown.
In short
Heard — Heard is a voice context layer that narrates your Claude Code and Codex agent sessions aloud so you can stop watching the terminal. Best for Mac developers running long Claude Code or Codex sessions who keep tab-switching to check progress, Engineers who step away from the desk and want iPhone text, voice-note, or call pickup on Power, Builders who want an open-source, inspectable voice layer with local Kokoro narration. Plans from $24/mo.
What is Heard?
Heard — Heard is a voice context layer that narrates your Claude Code and Codex agent sessions aloud so you can stop watching the terminal. Best for Mac developers running long Claude Code or Codex sessions who keep tab-switching to check progress, Engineers who step away from the desk and want iPhone text, voice-note, or call pickup on Power, Builders who want an open-source, inspectable voice layer with local Kokoro narration. Plans from $24/mo.
Heard is a macOS voice context layer for coding agents. Long Claude Code or Codex runs usually pin you to the terminal, waiting for a test to fail or a question to land. Heard distills those runs into short spoken updates, so you hear code changes, test outcomes, blockers, and open questions without reading logs. It sits on top of the agents you already use instead of replacing them. Narration is the free part. Claude Code and Codex CLI connect through first-class hooks; cloud agents like Grok Bot, Meta Muse, Devin, and Manus report through one MCP connector. Cursor, GitHub Copilot CLI, and Gemini CLI show sessions and approvals as status. It runs inside any terminal, including Ghostty, Herdr, cmux, iTerm2, and tmux. The open source Apache 2.0 engine can be built with local Kokoro voices or your own provider keys, so a fully private setup stays possible. The Power tier (paid) adds the interaction side. Push to Agent holds right Cmd to talk to any session, Parrot answers read-only questions about a session or codebase, and spoken recaps catch you up on demand. Give one voice instruction and Heard routes it to the right sessions, passing finished work between agents and running tasks in parallel. Away from the Mac, you can text, voice-note, or call Parrot from an iPhone. Privacy choices are explicit: source code, full terminal output, and agent transcripts stay on the Mac unless an enabled hosted feature sends a short text snippet; Heard says it does not retain narration text, audio, or transcripts on its servers. Against staring at a terminal or a generic text-to-speech script, the difference is structure: real agent events, a route back to the session that asked, and a phone companion for when you step away.
Key features
- Voice narration of agent progress, questions, blockers, and decisions
- Spoken test results and material state changes from coding agents
- First-class hooks for Claude Code and Codex CLI
- Cloud agents including Grok Bot, Meta Muse, Devin, and Manus via one MCP connector
- Cursor, GitHub Copilot CLI, and Gemini CLI show sessions and approvals as status
- Works in Ghostty, Herdr, cmux, iTerm2, and tmux
- Push to Agent: hold right Cmd to talk to any session (Power)
- Parrot voice conversations about sessions and code, read only (Power)
- Jump to a session or open apps by voice (Power)
- Spoken recaps on demand to catch up (Power)
- Pass finished work between agents and run tasks in parallel from one voice instruction
- Text, voice-note, or call Parrot from your iPhone (Power)
- Local Kokoro voice narration on macOS with no cloud call
- Bring your own AI and voice provider keys
- Open source Apache 2.0 engine with a heard run wrapper for other command-line agents
Behind the verdict
Most agent tooling fights for the same screen space. Heard goes the other way and assumes you want your eyes freed while the run grinds. We would reach for it the moment a single session runs past a few minutes and you notice yourself tab-flipping just to check whether tests passed. Pick it when your agents are Claude Code or Codex CLI, because those get first-class hooks and the events are structured. Cloud agents reporting through the MCP connector work, and status-only agents like Cursor and Gemini CLI are fine for awareness, but neither path is as rich as the hooked ones. If most of your work is one quick agent call, narration buys you little. The Power plan is where the product stops being a narrator and starts being a remote control. Holding right Cmd to answer a session, or asking Parrot to dig into a file while your hands are busy, is the part we keep coming back to. Pass it when you are on Windows or Android (macOS 14+ only), when your agents sit outside the supported list and you will not use the heard run wrapper, or in an open office where talking to a hotkey is socially awkward. Watch the daily narration ceiling on Power if you bank on continuous commentary across a large fleet. The engine being Apache 2.0 is an underrated fallback: with local Kokoro voices and your own keys you get a private setup without the managed tier. The closest alternative is not another AI tool so much as stitching generic text-to-speech onto your logs. That works until you need the route back to the session that asked, which is exactly the gap Heard fills.
Models under the hood
as of 2026-11-08
Limitations
- Heard is a macOS (14+) voice context layer; the iPhone companion is Power-tier only and your Mac must stay on, awake, and running the agents while you're away.
- The full voice loop (Parrot conversations on right ⌘, push-to-agent dictation, iPhone text/call) is Power-only; Free is the open-source summarization layer and requires bringing your own AI and voice keys.
- Managed narration on Power is capped at up to 2M characters a day.
- The microphone is not always on: it listens only while you talk to Parrot and for brief windows after replies and updates.
as of 2026-10-15
Researching Heard? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is Heard still active?
Heard is live · pricing from $24/mo · last verified 8 Nov 2026 · viability healthy.
Viability score
How well maintained and how widely used is Heard? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: November 2026
How we score →What do users say about Heard?
What people actually say about Heard — is it worth it?
Our community sentiment research written Jul 26, 2026 is more than 90 days old, so it is being re-reviewed against current data and is not shown.
How we verified this
Every listed fact about Heard — pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 8 Nov 2026. Analysis is drafted with AI models and checked against our editorial standards before publishing; see how we verify.
Verification history
We have re-verified Heard 10 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — changed SSL certificate
- — changed description, features, pricing tiers, who should skip it, our verdict, summary, pricing page, models, limitations
- — changed pricing tiers, who it suits, summary, integrations, features, who should skip it, our verdict, description
- — changed pricing page, limitations
- — changed SSL certificate
- — changed pricing tiers, who it suits, summary, description, who should skip it, our verdict, features, integrations
Showing the 6 most recent of 10 verification passes.
Free to cite with attribution — this page re-verifies continuously.
How much does Heard cost in 2026?
Heard: Plans from $24/mo. Pricing last verified 8 Nov 2026.
Plans compared
Every published Heard tier, its price as the vendor publishes it, who it fits, and what it adds over the tier below. The 12-month cost projection follows.
| Plan | Price | Billing | Who it's for |
|---|---|---|---|
| Free | Usage-basedVendor: “$0/forever” | usage | Mac developer who wants private narration and is comfortable supplying their own AI and voice provider keys with local Kokoro voices.Adds: The starting tier: runs locally, multiple agent sessions, local Kokoro narration, BYO keys, open source and inspectable. |
| Power | $24/mo | monthly | Engineer who steps away from the desk and wants to ask, answer, and direct agents by voice, including from an iPhone.Adds: Adds the microphone on top of Pro: doubles the daily limit to 2M characters, plus “Hey Parrot” conversations, voice session jumps, hold-right-⌘ dictation, ambient mode, and the iPhone companion. |
| Team | Custom | quote | — |
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Our hidden-costs note written Oct 15, 2026 quotes $12, $144, which no longer matches the verified price table, so it is being re-reviewed and is not shown.
Our price-fit note written Oct 15, 2026 quotes $12, $144, which no longer matches the verified price table, so it is being re-reviewed and is not shown.
Has Heard's pricing changed?
No plan price for Heard has moved since we began logging in Aug 2026; the pricing page is re-read on every verification pass.
Who should pick Heard?
Pick Heard if you are Mac developers running long Claude Code or Codex sessions who keep tab-switching to check progress, Engineers who step away from the desk and want iPhone text, voice-note, or call pickup on Power, Builders who want an open-source, inspectable voice layer with local Kokoro narration, Teams needing activity across projects, admin controls, and zero data retention on the Team tier.
Look elsewhere if you are Windows or Android users: Heard supports macOS 14 or later, with no Windows/Android today, Anyone whose agents fall outside the supported list and won't use the open source heard run wrapper, Silent shared offices where speaking to a hotkey isn't practical, Heavy always-on narration users who will brush the 2M daily character cap on Power.
Skip it if Skip Heard if you are on Windows or Android, your agents are outside the Claude Code, Codex, or MCP cloud-agent list and you won't build the open source engine, or your agent runs finish in seconds so narration overhead never pays for itself.
GeneratedReal-world workflow fit
Concrete scenarios for the personas Heard actually fits — and what changes day-one when you adopt it.
Kick off a multi-file refactor, pick up Pro, choose the Companion listening mode, and leave the desk. Heard narrates the blocker when the agent needs an auth decision and stays quiet through routine tool activity.
Outcome: You return only when the agent actually needs a decision, and the run finishes without you refreshing a tab every few minutes.
Open several sessions across Ghostty and tmux, let Jarvis narrate them into one stream, and ask for a spoken recap at the end of the day instead of scrolling terminal history.
Outcome: You can tell the sessions apart by ear — Heard ranks active ones and names the project when it switches — and you stop juggling windows.
Pair the iPhone companion, say “Hey Parrot” to ask about a session, and use hold-right-⌘ dictation or push-to-talk to answer an agent's question from the street.
Outcome: The build keeps moving while you're away from the desk, and you reply without opening the terminal at all.
Use cases
- Hear agent progress, blockers, and test results while your eyes are on something else
- Answer an agent's question mid-walk through iPhone listening and push-to-talk on Power
- Track multiple parallel Claude Code and Codex sessions by ear instead of tab-switching, with the project named when the session switches
- Get a spoken recap of what every agent accomplished without reading terminal scrollback
- Run other command-line agents through the same narration layer using the open source heard run wrapper
- Follow Codex app threads and reply into the composer without returning to the terminal
- Ask a website's underlying codebase what changed, with GitHub, Linear, and Jira context attached via Parrot
- Use voice instead of reading every terminal line if reading code output is a strain
Setup time & first value
How long it actually takes to get something useful out of Heard — broken out by persona, not the marketing-page minute.
Managed app: download, sign in, pick an integration, and hear your first event in minutes — the docs recommend verifying one agent with a small task before connecting more sessions. Open source engine: noticeably longer, since you clone the Apache 2.0 repo and supply your own AI and voice provider keys, or configure local Kokoro speech. Claude Code and Codex CLI hook setup is the fast path; cloud
Switching to or from Heard
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From watching the terminal: install the managed app, connect Claude Code or Codex CLI, and set a listening mode so you hear material changes instead of watching every line.
- →From a self-built open source setup: move to the managed Mac app to get hosted premium voices and personas without running your own provider keys.
- →From Cursor, Copilot CLI, or Gemini CLI only: keep those sessions visible in Heard for status and approvals, and connect Claude Code or Codex if you want actual narration.
- →From another command line agent: wrap it with the heard run wrapper from the Apache 2.0 engine to route it through the same narration layer.
- ↗To a plain terminal: nothing to unwind — Heard sits alongside your agents rather than replacing them, so stopping narration leaves Claude Code and Codex running as before.
- ↗To the open source engine: build the Apache 2.0 repo, bring your own AI and voice provider keys, and keep local Kokoro narration if you no longer want the managed services.
- ↗To a platform Heard doesn't cover: on Windows or Android you would need a different voice or notification layer entirely, since Heard requires macOS 14 or later.
Integrations
Resources & guides
- Documentationheard.dev
Docs · Heard
Full product docs from heard.dev
- Documentationheard.dev
Docs · Heard
Full product docs from heard.dev
- Documentationheard.dev
Claude Code · Heard
Full product docs from heard.dev
- Documentationheard.dev
Codex · Heard
Full product docs from heard.dev
- Quickstartheard.dev
Quickstart · Heard
Get up and running fast from heard.dev
Tutorials & Learning
YouTube returned 6 videos for “Heard”, and we withheld 6: 6 could not be judged, because “Heard” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Heard.
Official links
What are the best Heard alternatives?
The closest alternatives to Heard we list are Voicemode, fanbox and ADE. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| Voicemode | Free | Free | — |
| fanbox | Free | Free | — |
| ADE | Free | Free | — |
Alternatives to Heard
View all
Tools that pair well with Heard
Common stack mates teams adopt alongside Heard, with the specific reason each pairing earns its keep.
Voicemode
VoiceMode brings free, local voice conversation to Claude Code — speak coding tasks and hear answers back.
fanbox
FanBox is a free macOS terminal with a live diff viewer showing every file your Claude Code or Codex agent writes.
ADE
ADE runs Claude Code, Codex, Cursor and other AI coding agents side by side in one synced agentic development environment
Heard vs its closest rival
Heard's closest head-to-head is with Bito — Bito and Heard solve completely different problems. Bito is a must-have for engineering teams wrestling with multi-repo complexity — it turns architectural chaos into a live knowledge graph that your coding agents can query. Heard is a niche ergonomic tool for developers who want to voice-command their IDE. If you manage a large codebase and need better code quality across teams, choose Bito. If you type slowly or have RSI, Heard is a cheap productivity boost. There's little overlap: you could even use both. But for raw team efficiency, Bito delivers more tangible value. Verdict generated
Frequently asked questions
Our FAQ set (2 answers) written Oct 15, 2026 quotes $12, $144, which no longer matches the verified price table, so it is being re-reviewed and is not shown.
Categories
Topics
For makers of Heard
- Get a Verified Listing for Heard ($39): a written decision within 1 working day.
- Free listing for another tool: Free reviews are decided in order, within 3 weeks of the review starting.
Used Heard? Help shape our editorial sentiment research.