Skip to main content
Glama
alexbo-git

kalyvox-mcp

by alexbo-git
README.md
# Kalyvox MCP

Official public Model Context Protocol (MCP) server for **Kalyvox**, an AI phone receptionist for small businesses.

Kalyvox answers inbound calls 24/7 in natural language, identifies the caller's intent, captures structured information, routes or escalates requests, books appointments and sends actionable call summaries to the business.

**Website:** https://kalyvox.ai/en/  
**ChatGPT integration:** https://kalyvox.ai/en/integrations/chatgpt  
**Help center:** https://kalyvox.ai/en/help

## What this MCP server exposes

This repository contains a small, public, read-only MCP server designed for discovery and integration tooling. It exposes public information about Kalyvox without exposing customer accounts or internal APIs.

### Tools

| Tool | Purpose |
| --- | --- |
| `get_kalyvox_overview` | Get a concise overview of Kalyvox and its core use case. |
| `get_kalyvox_features` | List the main AI receptionist capabilities. |
| `get_kalyvox_integrations` | List the main supported integrations. |
| `get_kalyvox_links` | Return official Kalyvox product, help and integration links. |
| `find_kalyvox_help` | Route a setup or product question to the official Kalyvox help center. |

## Customer account data

Kalyvox also provides an authenticated ChatGPT integration for customers to query selected account and call data in natural language.

That authenticated integration is separate from this public repository. This repository **does not expose customer call data, transcripts, caller phone numbers, API tokens or private Kalyvox endpoints**.

Learn more: https://kalyvox.ai/en/integrations/chatgpt

## Install

```bash
npm install
npm run build
```

Run the MCP server over stdio:

```bash
npm start
```

Example MCP client configuration after building:

```json
{
  "mcpServers": {
    "kalyvox": {
      "command": "node",
      "args": ["/absolute/path/to/kalyvox-mcp/dist/index.js"]
    }
  }
}
```

## Development

```bash
npm install
npm run build
npm start
```

The server is implemented in TypeScript using the official Model Context Protocol TypeScript SDK.

## Kalyvox capabilities

Kalyvox is built for businesses that depend on inbound phone calls and need every request captured even when the team is busy or unavailable.

Core capabilities include:

- 24/7 AI call answering
- Natural-language voice conversations
- Caller intent detection and qualification
- Custom customer scenarios and priorities
- Structured call summaries and tickets
- Email and SMS alerts
- Human escalation and call transfer rules
- Google Calendar and Microsoft Outlook appointment booking
- Zapier automation
- ChatGPT integration for account and call analysis
- Multilingual call handling

## Security

This public MCP server contains no Kalyvox production credentials and requires no Kalyvox customer token. Account-specific functionality remains behind Kalyvox authentication and user-controlled sharing settings.

## MCP Registry

Registry identifier:

```text
io.github.alexbo-git/kalyvox-mcp
```

The included `server.json` is prepared for MCP Registry publication once the npm package is published.

<!-- mcp-name: io.github.alexbo-git/kalyvox-mcp -->

## License

MIT © 2026 KALYVOX SASU

TDQS

A3.6/5.0

Scored across 5 tools

Disambiguation3/5

Overview and features overlap conceptually, and both get_kalyvox_links and find_kalyvox_help can return help-center URLs. The descriptions provide enough distinction for most use cases, but some misselection is possible.

Naming Consistency4/5

Tool names follow a consistent get_kalyvox_* pattern with clear noun suffixes. find_kalyvox_help is the only deviation, but it still feels natural and readable.

Tool Count5/5

Five tools is well-scoped for a small informational product server. Each tool covers a distinct area of product information without unnecessary duplication.

Completeness4/5

The set covers overview, features, integrations, links, and help routing, which addresses the main informational needs for an AI phone receptionist product. Minor gaps such as use cases, pricing, or FAQ content are not exposed, but agents can work around them via the help-center tools.

Maintenance

ActivityMaintained
ResponsivenessUnresponsive