three.ws Concierge runs on your own machine — the client starts it, so there is no endpoint to ping. 204 installs a week from npm. Last commit 18 Sep 2026.
Ask any website's AI concierge a grounded question, and generate the embed to add one to a site.
Today is the operative word: we check three.ws Concierge every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add concierge-mcp -- npx -y @three-ws/concierge-mcp
{
"mcpServers": {
"concierge-mcp": {
"args": [
"-y",
"@three-ws/concierge-mcp"
],
"command": "npx"
}
}
}
[mcp_servers.concierge-mcp]
command = "npx"
args = ["-y", "@three-ws/concierge-mcp"]
{
"mcpServers": {
"concierge-mcp": {
"args": [
"-y",
"@three-ws/concierge-mcp"
],
"command": "npx"
}
}
}
{
"mcpServers": {
"concierge-mcp": {
"args": [
"-y",
"@three-ws/concierge-mcp"
],
"command": "npx"
}
}
}
This one needs environment variables set before it will start:
THREE_WS_BASE (Base URL of the three.ws API that serves POST /api/concierge, and the origin the generated embed snippets point at. Override only when self-hosting or targeting a preview deployment.), THREE_WS_TIMEOUT_MS (Per-request timeout in milliseconds for the concierge answer call. The answer streams over a free-first LLM chain that can fail over across providers, so the default leaves room for one slow rung.), CONCIERGE_PAGE_TIMEOUT_MS (Timeout in milliseconds for fetching a page that concierge_ask grounds its answer in. Separate from the answer timeout so a slow site cannot consume the whole call.).
The author declared them in the registry entry; get the values from the project itself.
Generate a paste-ready embed for a floating 3D avatar chatbot on any website.
IBM watsonx.ai MCP by three.ws: chat, text generation, embeddings, and tokenization.
Create AI-generated exam questions and save them to your paperee question bank for review.
Ask a codebase questions and get cited code back, from a tree-sitter graph of the repo.
Search the web, generate AI responses, and create embeddings with real-time information
The ask-and-answer video format: send a question, get a short one-take video answer, captioned.
MCP server exposing OptiQra's website audit tools: SEO, GEO, AEO, and AI-generated fixes.
The website platform for AI agents. One API to build, host, and operate real websites.
Answers built from our own checks of this server.