Pricing
These are the planned tiers for when Speech Layer opens. Your audio stays on your machine on every tier. Join the waitlist for launch notice — checkout is not the door yet.
Community
The full speech layer on your machine
or $2000/yr billed annually when billing opens
Join the waitlist- Full local speech layer: wake word, speech-to-text, text-to-speech
- Voice agents and multi-worker orchestration
- Browser automation (surfagent skill)
- Unlimited local use on Linux today, macOS coming
- Community support
Pro
Convenience for daily drivers
or $4000/yr billed annually when billing opens
Join the waitlist- Everything in Community
- Premium bundled voices
- Unlimited voice cloning, created and stored locally
- Optional cloud STT/TTS fallback for speed (opt-in, off by default)
- Early access to new agents and skills
- Priority support
Team
One speech layer for the whole crew
or $5000/user/yr billed annually when billing opens
Join the waitlist- Everything in Pro
- Shared agents, macros, and voice commands across a team
- Synced configs across machines
- Central management and admin controls
- Team support
Prices in USD, planned for when paid tiers open. Every plan keeps speech on your machine. Release expected September 2026 — join the waitlist for notice. Developer preview continues on Linux; macOS is coming.
What you get at each level
The core engine is identical on every tier. Higher plans add convenience, polish, and team features on top.
| Feature | Community | Pro | Team |
|---|---|---|---|
| Wake word, local STT and TTS | Yes | Yes | Yes |
| Voice agents and orchestration | Yes | Yes | Yes |
| Browser automation (surfagent) | Yes | Yes | Yes |
| Linux today, macOS coming, unlimited local use | Yes | Yes | Yes |
| Audio stays on your machine | Yes | Yes | Yes |
| Premium bundled voices | — | Yes | Yes |
| Unlimited voice cloning (local) | — | Yes | Yes |
| Cloud STT/TTS fallback (opt-in) | — | Yes | Yes |
| Early access to agents and skills | — | Yes | Yes |
| Shared agents, macros, and commands | — | — | Yes |
| Synced configs across machines | — | — | Yes |
| Central management / admin controls | — | — | Yes |
| Support | Community | Priority | Team |
Questions, answered
Is my audio really private on every tier?
Yes. Wake word detection, transcription, and speech synthesis all run locally on every tier. Higher tiers add convenience features, not data collection. We never see or store your audio.
How does voice cloning work, and where are my clones stored?
Pro includes unlimited voice cloning: record a short sample and Speech Layer builds a voice you can use for any agent or spoken reply. Clones are created and stored entirely on your machine. Like everything else in Speech Layer, your voice never leaves your hardware unless you explicitly opt into a cloud feature.
What is the cloud STT/TTS fallback in Pro?
An optional speed boost: if you enable it, a transcription or voice request can go to a cloud model when you want the fastest possible response. It is strictly opt-in, off by default, and clearly indicated in the UI when active. Turn it off and nothing ever leaves your machine.
Which platforms are supported?
Linux on every tier today, with macOS support on the way. Speech Layer is built for real desktops: PulseAudio and PipeWire on Linux, Core Audio on macOS when it lands, with packages for the major distributions.
Can I cancel anytime?
Yes. There are no contracts and no lock-in. Cancel in a couple of clicks and your plan simply does not renew. Your local install and history stay on your machine either way.
What happens to my local data if I cancel?
Nothing. Your configs, voices, agents, and history live on your machine, not in our cloud. Cancelling a subscription changes nothing about the files already on your hardware.
Be on the list for launch
We will email you when Speech Layer is ready. Release expected September 2026. Tiers above are the planned ladder, not a buy-now checkout.