About handset
handset provides open-source React components and hooks for embedding SMS messaging in a product using the Handset API. Its components are copied into the project as editable source, while a backend proxy keeps the API key off the client.
Key features
- 01Responsive inbox and message thread components
- 02Composer with SMS segment and Unicode detection
- 03React hooks for conversations, threads, and composition
- 04Next.js proxy routes for API requests
- 05Source components styled with project Tailwind tokens
Use cases
- 01Add an SMS inbox and thread to a product
- 02Build a custom messaging interface with React hooks
- 03Keep the Handset API key on the backend
- 04Show SMS segment counts while composing
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y @handset/mcpHANDSET_API_KEYSecretHandset API key. Use a test-mode key (hs_test_…) — simulated carrier, free. Live keys are refused unless HANDSET_ALLOW_LIVE=1.HANDSET_ALLOW_LIVESet to 1 to permit a live-mode key (the agent can then text and call real people at real cost).Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
15Saved tool list · updated 2w ago
| Parameter | Type | Required | Description |
|---|---|---|---|
| limit | integer | No |
Details
From the project READMEHandset UI
Texting inside your product, in an afternoon.
Open-source React components for embedded messaging, built on the Handset API. Distributed the shadcn way: the CLI copies the source into your repo, styled with your Tailwind tokens, yours to change.
# 1. Point the shadcn CLI at the Handset registry
npx shadcn@latest registry add @handset=https://ui.handset.dev/r/{name}.json
# 2. Install a complete texting surface + the server proxy route
npx shadcn@latest add @handset/messaging @handset/next-routes
Set HANDSET_API_KEY in your env, wire resolveTenantId() in the generated
route to your session, and render it:
import { HandsetProvider } from "@handset/react";
import { Messaging } from "@/components/handset/messaging";
export default function MessagesPage() {
return (
<HandsetProvider>
<div className="h-[600px]">
<Messaging />
</div>
</HandsetProvider>
);
}
What's in the registry
| Item | What it is |
|---|---|
@handset/messaging | Two-pane texting surface — inbox + thread, responsive to mobile |
@handset/inbox | Conversation list: previews, relative times, STOP badges, infinite scroll |
@handset/thread | Message history: delivery states, opt-out handling, wired composer |
@handset/composer | Draft box with live SMS segment counting and unicode detection |
@handset/next-routes | Next.js proxy route — your API key never reaches the browser |
Architecture
Two layers, like shadcn-on-Radix:
@handset/react(npm, versioned) — headless hooks:useConversations,useThread,useComposer, plus segment math and a tiny client. Logic we maintain; bring your own UI if you prefer.- The registry (
apps/www/registry) — the visual components, copied into your project as source.
Components never hold API keys. They call proxy routes in your backend
(added by @handset/next-routes); your server holds the Handset key and your
existing session auth decides which tenant each signed-in user sees. The
Handset API handles the rest — numbers, delivery, threading, 10DLC compliance,
STOP/HELP.
Development
pnpm install
pnpm --filter @handset/react build # hooks package
pnpm dev # docs site + live demo at localhost:3000
pnpm registry:build # compile registry.json → public/r/*.json
The docs site serves a mock Handset API (apps/www/app/api/handset) so the
demo works with zero setup.
Requirements
- React 18+ (components target React 19 / Tailwind v4 / shadcn CLI ≥ 3)
- A Handset account for real traffic — the demo runs on mock data without one
Roadmap
See ROADMAP.md — shipped components, what's next (call summaries, MMS attachments, more framework proxies), and non-goals. Issues with real use cases move items up the list.
License
MIT © Handset HQ, Inc.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.