command-bridge

Talk to your coding agent from your phone. One command opens a page any phone browser can load — no app, no App Store — and carries audio both ways.

I wanted my own Jarvis: to drive my coding agents by voice while out of the house, without handing my screen or my speech to a third party. So the premise is deliberately small — a local CLI, so if your agent can run bash, it can talk to you. No server to host, no account, no API key.

Speech recognition and speech synthesis both run on your own machine, and the CLI is fully self-describing: your agent installs it and immediately knows how to drive it, without an MCP server or a doc you have to keep in sync. One caveat, and it is a browser rule rather than a choice: to use this from outside your own network you will want Tailscale or ngrok, because a phone will not give a web page a microphone over plain http.

Watch it work

Talking to a team of Claude Code agents by voice: ask for the roster, bring agents up, see the board of what each one is on, and watch one explain a blog post with a diagram it points at box by box.

Quick start

One command installs it straight from this repo — no npm, no PyPI, no account:

curl -fsSL https://raw.githubusercontent.com/JuanjoFuchs/command-bridge/main/install.sh | bash

On Windows, in PowerShell:

irm https://raw.githubusercontent.com/JuanjoFuchs/command-bridge/main/install.ps1 | iex

It finds Python, clones the repo into ~/command-bridge, builds a private virtualenv beside it, and installs command-bridge with its neural engines — nothing touches your system Python. Put <repo>/bin on your PATH and command-bridge works from anywhere. Then paste this to your coding agent:

Run command-bridge describe and follow it end to end: install anything missing, start command-bridge under your own name, give me the URL to open on my phone, and then stay in watch so you can hear me.

That is the whole handoff. Your agent downloads the models, starts the server and hands you back a URL. Open it on your phone, tap once, and say "hey Claude, can you hear me?"

Needs Python 3.10+ on PATH. The installer builds a private environment inside the checkout and touches nothing else on your machine — removing that one directory removes all of it.

The one thing your agent cannot do for you: a phone needs HTTPS to reach a microphone at all, so a plain LAN address gives no microphone rather than a broken one. Front the port with a tunnel — ngrok http 8765 is the smallest, and command-bridge status detects it — and use the https address it prints. Your agent will tell you when this applies, and status.phone.exposure tells you what it costs.

Everything runs on your machine. No GPU. No account. No speech API.

Would rather drive it yourself? Use it is the manual path.

Why one line is enough

Point the agent at describe before anything else — the line above does exactly that:

command-bridge describe     # the whole contract, as JSON

That one call is the entire onboarding — no MCP server, no daemon, no separate documentation to keep in sync. It returns the loop to run, the turn schema, every command and argument, the exit codes, and the two rules an agent gets wrong first: watch blocks, so an agent not sitting in it has left you talking to nobody; and one thought arrives as several turns, so answering the first one answers the wrong question. Every command also returns a next field telling the agent what to do at the moment it applies, rather than in a document it read once.

describe is generated from the same source as the behaviour, so it cannot drift from the tool the way a README can. When something is wrong, command-bridge doctor says what and hands back the command that fixes it — read its degraded list even when ok is true, because a machine can have a neural voice and a fast recognizer sitting on disk while this process uses neither.

That claim gets tested the only way it can be: hand an agent a fresh install, nothing but command-bridge describe, and ask it to reach the best configuration the machine allows. Several releases have come out of watching where those runs got stuck.

The wedge is borrowed from agent-mail.

What this is

The tool holds no model and makes no decisions. It turns your speech into lines in a log and text into speech; the agent that started it does the thinking. That is why it works with any agent — Claude Code, Codex and Grok have each driven it unchanged.

The cursor is what makes the split safe. watch --since <cursor> blocks until a turn lands and returns everything after the cursor, so nothing you said is dropped while the agent spends thirty seconds thinking about the last thing.

Install

Published packages (npm, PyPI, WinGet) are not out yet — for now command-bridge installs straight from this repository.

One line

curl -fsSL https://raw.githubusercontent.com/JuanjoFuchs/command-bridge/main/install.sh | bash   # macOS / Linux

irm https://raw.githubusercontent.com/JuanjoFuchs/command-bridge/main/install.ps1 | iex          # Windows

The installer finds Python 3.10+, clones the repo (into ~/command-bridge, or wherever CB_DIR points), builds a private virtualenv beside it, and installs the package with every engine. It runs the same whether piped or from a checkout (./install.sh / .\install.ps1), and reruns cleanly to update. Nothing touches your system Python, and deleting the checkout removes all of it.

By hand

git clone https://github.com/JuanjoFuchs/command-bridge.git
cd command-bridge
python -m venv venv
venv/bin/python -m pip install -e ".[all]"      # venv/Scripts/python.exe on Windows
export PATH="$PWD/bin:$PATH"                     # then `command-bridge` works from anywhere

Better voice and better recognition

The one-line install already includes every engine. command-bridge setup then downloads all of the models in one command — or fetch them individually:

command-bridge download asr          # Parakeet — 8x faster than whisper, more accurate
command-bridge download voice        # a neural voice instead of the robotic one
command-bridge download voiceprint   # learns your voice, so the wake phrase becomes optional
command-bridge download turn         # ends your turn when you SOUND finished, not on a timer
command-bridge download --list       # what is available, what you already have

Two independent things are involved and having one does not get you the other: the extras supply the engines, the downloads supply the models. command-bridge doctor says which of them this process is actually using, which is not always the best one present.

Models are downloaded, never bundled — a Parakeet checkpoint is 631 MB and would make the package unusable on a slow connection.

Use it

The manual path, if you would rather not hand the whole thing to an agent. Start command-bridge with your agent's own name, and put its page somewhere your phone can reach.

command-bridge serve --wake claude          # or codex, grok, whatever is driving

Your phone needs HTTPS to reach a microphone at all — that is a browser rule, and a plain LAN address like http://192.168.1.20:8765 gives no microphone, not a broken one. Front the port with an https tunnel. The smallest is ngrok, which command-bridge status detects on its own:

ngrok http 8765

tailscale serve --bg 8765 works too and needs command-bridge config set COMMAND_BRIDGE_ALLOW_CIDRS 100.64.0.0/10 — but it takes over the device's DNS through MagicDNS, system-wide, which can break a corporate VPN on the same machine. Anything else (cloudflared, a reverse proxy) is invisible to this tool, so tell it: command-bridge config set COMMAND_BRIDGE_PUBLIC_URL <https url>.

Read this before you open a tunnel. A forwarder connects from 127.0.0.1, so everything it relays arrives as a loopback peer and passes COMMAND_BRIDGE_ALLOW_CIDRS unconditionally. While a tunnel is up that allowlist filters nothing, and the token in the URL is the only gate on a live microphone — so treat that URL as a credential, set COMMAND_BRIDGE_TOKEN yourself rather than letting it be regenerated each restart, and take the tunnel down when you are done. command-bridge status reports all of this under phone.exposure; command-bridge doctor warns when the only gate is a token nobody chose.

Then, from the agent's side, this is the whole loop:

command-bridge watch --session dev --since -1   # BLOCKS until you speak; returns turns + a cursor
command-bridge say   --session dev "All green." # speaks back
command-bridge watch --session dev --since 7    # always resume from the cursor you were given

Saying its name

The summons is a greeting plus a name: "hey claude", "hi codex", "ok grok". The greeting is always required, which is what makes any name safe — grok is an English verb and cursor is a word you say constantly, but nobody says "hey grok" by accident.

command-bridge wake --name codex     # change it, live, no restart

Once the voiceprint knows you, you are addressed without saying anything.

Interrupting it

Start talking while it is speaking and it stops — but only for your voice. The voiceprint has to agree before a reply is cut off, so the television, someone else in the room, and the agent's own voice coming back through your speakers all leave it talking.

That last one is not a corner case. Without echo cancellation a reply leaks into the microphone on almost any device, and "the agent interrupts itself" is the default failure. Measured here: the owner's voice scores 0.23 against a 0.15 threshold, the agent's own voice scores 0.000.

Needs command-bridge download voiceprint and a voice it has learned. Without one it stays quiet rather than guessing, because a bridge that stops whenever the room makes a noise is worse than one you cannot interrupt.

Knowing when you have finished

By default a turn ends after a fixed silence — 1.5 seconds. That number is a compromise: shorter cuts you off while you are still thinking, longer makes every quick question wait for nothing.

command-bridge download turn replaces it with Smart Turn v3.2 (8 MB, CPU, BSD-2-Clause), which listens to how the sentence ended. A finished question closes early; a trailing "I was thinking that maybe we could…" gets more room. It runs once at each pause, not continuously, and if it is not installed the timer works exactly as it always has.

Requirements

No GPU. Every model here is CPU inference by construction — nothing in the codebase can use a graphics card. Measured on a 20-core desktop CPU:

• Python 3.10+ on PATH.
• A phone browser. Android Chrome is what this is tested on. The tab must stay in the foreground — background recording needs a native app, which this deliberately is not.
• HTTPS to the phone, via Tailscale or any tunnel. See Privacy.
• A slower CPU raises the real-time factor but does not break anything; the recognizer runs faster than real time with a lot of headroom.

Privacy

Speech recognition, synthesis, the voiceprint, and every turn of the transcript stay on your machine. There is no account and nothing is sent to a speech API.

The transport is your choice and they are not equivalent. Tailscale terminates TLS on your own device, so nobody else can decrypt the audio. An ngrok or Cloudflare tunnel terminates it at the vendor's edge — which puts the plaintext of everything you say on a machine you do not control. Sometimes that is the only thing that works, and it is still a trade worth making knowingly. If you expose it publicly, put real authentication in front of it: the built-in CIDR allowlist cannot help you there, because a tunnel forwards from localhost and every request therefore arrives from an allowed peer.

Your files are plain files. command-bridge config path says where. The voiceprint is a 192-dimension centroid — speech cannot be reconstructed from it.

Settings

command-bridge config show           # every setting, its value, and where it came from
command-bridge rate --speed 1.4      # talk faster — applies now and every session after
command-bridge verbose on            # narrate every action before doing it
command-bridge timing                # where the time actually went, per exchange

Precedence is process env > settings file > built-in default. .env.example documents every variable.

Development

python -m venv venv
venv/Scripts/python -m pip install -e ".[dev,all]"

venv/Scripts/python -m pytest tests/ -q     # unit tests, no mic and no model
venv/Scripts/python scripts/e2e.py          # the pipeline, through a real browser
venv/Scripts/python scripts/layout.py       # page geometry, at five viewports
venv/Scripts/python scripts/channel.py      # the orb, mute and the speaking signal
venv/Scripts/python scripts/diagram.py      # regenerate the diagram above, both themes

License

MIT.