Listening
Earshot gives your machine ears and a mouth. Speak into anything, hear your agents answer, and your voice never leaves the room.
Every way into a computer has been yours outright. The punch card, the command line, the trackpad, the finger. Bought once, working in a blackout, nobody counting. Then the most human one of all arrived metered, and listening somewhere else.
Speech is a hundred thousand years older than writing. It should not require a subscription.
Four suppliers, two of them cheap and uncertified. It does not guess which you meant and it does not order first and tell you after. It asks, out loud, and waits.
Talking to an agent all day means saying half-formed things out loud, the ones you have not told anyone yet. Recognition and synthesis both run on your Mac, so the recording never becomes anybody else’s to keep.
There is no account to sign in to, and nothing it sends has anything to do with what you said.
Every clip below is the app running. None of it is a mockup.
Free is one person and one agent, whole, with nothing switched off and no clock running. Pro is for when the room fills up, and Team is when the room is other people.
One person, one agent. No time limit.
Apple silicon, macOS 12 or later.
Once there is more than one agent, session, or machine.
One Mac, and the key is yours to keep. Each Mac checks in about once a month.
Everyone on the team, sharing vocabulary and config.
One seat per Mac, and the key is yours to keep. More than five hundred? Talk to us.
| Feature | Free | Pro | Team |
|---|---|---|---|
| Dictation into any application | Included | Included | Included |
| Speaking, asking and status, for any agent | Included | Included | Included |
| Every voice in the catalogue | Included | Included | Included |
| Your own shortcut | Included | Included | Included |
| Vocabulary corrections | Included | Included | Included |
| The last two hundred transcripts | Included | Included | Included |
| Runs with no account | Included | Included | Included |
| Needs the network | To fetch models and updates | The same, and once a month for the seat | The same, once a month per Mac |
| Hands-free, and “Hey Earshot” | Not included | Included | Included |
| A voice for each project | Not included | Included | Included |
| Muting one project while it works | Not included | Included | Included |
| Spoken answers matched to the agent’s options | Not included | Included | Included |
| Which session is talking, when one agent has two | Not included | Included | Included |
| Macs | As many as you like | One | One per seat |
| Billing | None | Once | One invoice |
The wake word in here is ours. We trained it because every pre-trained one worth having is licensed for non-commercial use, and we made recognition and synthesis small enough to run on a laptop. That is the part of Earshot we are least done with: more of our own models, tuned for agents and for the way people actually talk to machines. If your product needs the same, bring us the problem.