teamyavio/yavio

MCP应用和ChatGPT应用的产品分析

开源。自托管。看看用户是如何实际使用你的AI工具的。

---

Alpha（v0.1） — Yavio 正在积极开发中。API、配置和数据库模式可能会在不同版本之间发生变化，无需预告或迁移路径。在生产环境中使用时，除非钉定精确版本。

什么是雅维奥？

Yavio 是 MCP 服务器和 ChatGPT 应用的分析与可视化层。它通过三行 SDK 集成和完整的分析仪表盘，捕捉了 AI 平台（ChatGPT、Claude、Cursor、Windsurf、VS Code 等）如何调用你的工具，以及用户如何与你的小部件互动——工具调用、转换、漏斗、错误、保留——并配有三行 SDK 集成和完整的分析仪表盘。

整个平台都是**开源（MIT）**的，并通过Docker Compose自托管。云托管即将到来。

快速入门（自主持）

1.启动平台

仪表盘位于http://localhost:3000，摄取API在http://localhost:3001。

2.创建一个项目

打开仪表盘，注册账户，创建工作区和项目，然后复制你的 API 密钥。

3.安装 SDK

4.封装你的MCP服务器

就这样。

Product Analytics for MCP Apps and ChatGPT Apps

Open-source. Self-hosted. See how users actually use your AI tools.

---

Alpha (v0.1) — Yavio is under active development. APIs, configuration, and database schemas may change between releases without notice or migration paths. Do not use in production without pinning exact versions.

What is Yavio?

Yavio is the analytics and visibility layer for MCP servers and ChatGPT Apps. It captures how AI platforms (ChatGPT, Claude, Cursor, Windsurf, VS Code, etc.) invoke your tools and how users interact with your widgets — tool calls, conversions, funnels, errors, retention — with a 3-line SDK integration and a full analytics dashboard.

The entire platform is open source (MIT) and self-hosted via Docker Compose. Cloud hosting is coming soon.

Quick Start (Self-Hosted)

1. Start the platform

git clone https://github.com/teamyavio/yavio.git
cd yavio

# Generate .env with random secrets
./scripts/setup-env.sh

# Start all services (migrations run automatically on first start)
docker compose up -d

The dashboard is at http://localhost:3000, ingestion API at http://localhost:3001.

2. Create a project

Open the dashboard, register an account, create a workspace and project, then copy your API key.

3. Install the SDK

npm install @yavio/sdk

4. Wrap your MCP server

import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { withYavio } from "@yavio/sdk";

const server = withYavio(
  new McpServer({ name: "my-app", version: "1.0.0" }),
  {
    apiKey: "yav_your_key_here",
    endpoint: "http://localhost:3001/v1/events",
  }
);

That’s it. Every tool call is now tracked automatically.

Using the CLI

You can also use the CLI for setup and platform management:

npm i -g @yavio/cli

# Interactive SDK setup — creates .yaviorc.json
yavio init

# Platform management
yavio up          # Start services
yavio down        # Stop services (data preserved)
yavio status      # Health check and stats
yavio logs        # Tail service logs
yavio doctor      # Diagnose setup issues

Features

• Zero-config MCP instrumentation — withYavio() wraps your MCP server and captures every tool call, latency, errors, and platform automatically