Join the waitlist · release expected August 2026

Join waitlist
Speech Layer logo Speech Layer Developer preview · Alpha
Release expected August 2026 · Join the waitlist

A simple & fun way to drive
parallel swarms of AI agents
using only your voice.

 

Speech Layer brings conversational, local-first, massively parallel AI agent swarm control to Linux, with macOS support coming.

Your words are transcribed on your machine and routed to AI agent swarms (Pi via herdr) to work in parallel. They'll talk you through it in a distinct voice per project. You can shush or interrupt them any time.

Offline by design, works with llama.cpp or bring your OpenRouter API key. Achieve very real, insanely productive work hands-free and screen-free.

Join the waitlist Preview the install path
The fastest way to understand Speech Layer

Have a conversation with your computer

Press play and listen: Speech Layer hears the request, puts agents to work on it, often several at once, and speaks the answer back. Six short examples below.

    Loading demo…
    0:000:00
    Built for people who code

    Code by voice. Agents fan out and report back.

    Speak a task mid-flow: it is transcribed on your own machine, fanned out to coding agents that work in parallel, and when they finish your machine reads the summary back. Hands free, private by design.

    “Fix the login redirect, then run the tests”
    fanned out, in parallel
    worker: fix-login
    patch ready
    worker: tests
    running: 38 of 41
    worker: review
    reading the diff
    worker: changelog
    entry drafted
    layer says: “Fix is in, 41 tests green, diff is ready for review.”
    What it does

    Speech in, action out

    A thin, dependable layer between your voice and your toolchain. No cloud round trips required.

    Always-ready wake word

    Say the wake word mid-flow and start dictating: no window switching, no hands off the home row.

    Local speech-to-text

    Whisper-class transcription runs on your machine. Dictate a commit message or a whole feature brief, online or off.

    Agents that code

    Intents fan out to scripts and coding agents that do the work: dispatch a fix, run the tests, review a diff. Mail and browsing come along too.

    Spoken confirmation

    Fast local text-to-speech reads back summaries and asks the next question, so you steer a whole fleet of workers without touching the keyboard.

    Why Speech Layer

    The voice plumbing your desktop has been missing

    Speech Layer treats speech as system plumbing: boring in the best way, and always there when you reach for it.

    Private by design

    Audio capture, wake word detection, and transcription happen on your hardware. Your code and your voice never leave the machine unless you ask them to.

    Local-first

    Core features keep working with the network cable pulled. Cloud models are an option you opt into, never a requirement.

    Linux native, macOS coming

    First-class support for real Linux desktops today: native audio, sensible packaging, and docs for the systems people actually run. macOS is next.

    Scriptable and agent-ready

    Every intent is just a command. Wire voice to git, your test runner, your multiplexer, or full coding agents with a few lines of config.

    Supported platforms

    Runs where you work

    Speech Layer runs on Linux desktops today, with packages and docs for the distributions people actually use. macOS support is coming.

    LinuxLinux
    macOSmacOSComing
    UbuntuUbuntu FedoraFedora DebianDebian Arch LinuxArch Artix LinuxArtix NixOSNixOS

    Also tested on openSUSE Tumbleweed. Other systemd or OpenRC based distributions generally work via the generic installer.

    Waitlist

    Be first when it ships

    Release expected August 2026. Join the waitlist for a launch notice. The install path below is a developer preview for people already poking at the stack, not the main door.

    Join the waitlist Preview the install path 
    # Developer preview only (Linux; macOS coming)
# Release expected August 2026 — join /waitlist/
$ curl -fsSL https://speechlayer.com/install.sh | sh

# check the daemon
$ speechlayer status

# say the wake word and go
$ speechlayer start
    Under the hood

    Built like infrastructure, not a demo

    Speech Layer is the voice front end of a real agent stack. You speak, a local transcriber turns it into text, herdr fans the work out to panes of pi agents, and the result comes back as speech. Every layer is inspectable, scriptable, and replaceable.

    Whisper hears you

    Whisper-class transcription runs on your machine and turns your voice into text in well under a second. Nothing leaves the box unless you ask it to.

    herdr multiplexes agents

    Speech Layer drives herdr, the terminal multiplexer. One spoken request becomes named panes on a workspace: each an isolated worker with its own context, visible on screen and scriptable from anywhere.

    pi runs the harness

    Every pane runs an instance of pi, the coding-agent harness used by projects like OpenClaw. Ask for three workers and three agents run in parallel; Speech Layer merges what they report and reads the answer back.

    The loop

    One loop, five stages

    Each stage is a separate, replaceable component. Swap the transcriber, the intent router, or the voice, the pipeline stays the same.

    Wake word

    Low-power detector listens for your chosen phrase and opens the mic only then.

    Speech-to-text

    On-device transcription turns your request into text in well under a second.

    Intent and agent

    The router matches what you said to a script, tool, or AI agent with the right context.

    Action

    The agent does the work: reads mail, moves windows, runs builds, controls the desktop.

    Spoken reply

    A fast local voice confirms what happened, so you never break flow to check a screen.