Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@sidevoice-connector join the voice call so I can hear your progress"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Reading your coding agent's plans, diffs and summaries all day is tiring. Sidevoice turns the conversation you already have with your agent into a voice call. The agent keeps its context and keeps writing as usual; it also speaks its replies, and you answer by voice and can interrupt it — from the sofa or on a walk, not only at your desk.
sidevoice-connector is what you install on the machine where your agents run, published on npm as
@sidevoice/uplink. It gives each agent conversation its voice
tools (an MCP server), delivers what you say into that same conversation, and installs, starts and supervises the
machine's core.
Your machine is the computer where your coding agents run; a device is what you call from (the desktop app, a browser).
Piece | Role |
sidevoice-connector (this repository) | Installed next to your agents: their voice tools and the core's supervisor. |
The conversations and the voice pipeline, next to the agents. | |
The app you call from. | |
The call interface the app bundles; it can also be served as a static site. |
Related MCP server: livechat-mcp
You need Node.js 22 or newer and uv (the core runs on Python, which uv provides).
Install, on the machine where your agents run:
npx -y @sidevoice/uplink installIt installs and starts this machine's core, registers the voice tools with the agents it finds, and tells you what, if anything, needs a manual step (see Supported agents).
Pair your device. Ask your agent to pair a device, or run npx @sidevoice/uplink pair-device. You get a
one-time code; enter it in the desktop app. Pairing is always
your act: nothing pairs on its own.
Talk. In any conversation, ask the agent to join the voice call. It shows up in the app; speak to it.
Agent | Setup | What you say reaches the conversation |
Claude Code | Registered by | Delivered into the running session. |
Codex |
| Queued into the thread with |
Cursor (experimental) | Registered by | Cursor offers no route of its own. The CLI receives it only in a chat started with |
Any other agent can receive through an HTTP endpoint of its own (SIDEVOICE_DELIVERY_URL).
Beta. What works today: the install above on macOS and Linux, pairing devices with the machine, voice calls with your agents' conversations from the desktop app, interrupting a reply, and replies spoken while the agent keeps working. Reaching your machine from outside your network needs a relay, which is still being built.
The package installs one command, sidevoice:
Command | What it does |
| Installs this machine's core, starts it, checks it answers, then registers the voice tools. |
| The reverse of |
| Prints a one-time code (and its QR) that pairs a device with this machine. |
| The MCP server an agent starts, one per conversation. You do not run it yourself. |
| The per-machine process the MCP servers share; started on demand, gone when the last conversation leaves. |
packages/connector/README.md describes how the pieces talk to each other.
npm ci
npm test # node --test, on the source
npm run build # the bundled dist/cli.mjs that npm publishesThe published package has no runtime dependencies: esbuild bundles everything into dist/cli.mjs.
SIDEVOICE_CORE_WHEEL=<wheel> npm run build embeds the pinned core's wheel; installing it still lets uv download
Python and the core's dependencies.
Issues and pull requests are welcome. Read AGENTS.md first: it holds the rules for code, texts and
tests, for people and coding agents alike. Pull request titles follow
Conventional Commits (CI checks them) and become the squashed commit, from
which release notes are written (RELEASING.md).
Apache-2.0. The Sidevoice name and logo are trademarks: forks are welcome under their own name — see
TRADEMARKS.md.
The bundle redistributes its dependencies under their own licences: THIRD_PARTY_NOTICES.md.
This server cannot be deployed
Voice and chat for AI agents — Discord, Teams, Meet, Slack, Zoom, Telegram, WhatsApp, NC Talk, SIP
Give AI agents a phone layer for consent-based calls, transcripts, summaries, and outcomes.
Give your AI agent a phone: place calls, navigate IVRs, wait on hold, get structured answers.
Audio for your agent: transcribe, speak, translate, summarise, plus sound effects and music.
