yawa
Self-hosted web analytics platform that exposes analytics data via the Model Context Protocol, enabling natural language queries about traffic, page views, web vitals, and custom events.
README
yawa - Yet Another Web Analytics
[!IMPORTANT]
yawa is in early development. APIs, configuration, and data formats may change without notice. Use in production at your own risk.
yawa is a self-hosted web analytics platform with no UI. Instead of dashboards, you query your data through the Model Context Protocol (MCP), asking Claude or any compatible AI client questions about your traffic in plain language.
"List my analytics sites"
"What are my top pages this week?"
"Show me web vitals for my homepage."
"Which UTM campaigns are driving the most visitors?"
It, obviously, requires no cookie banners or GDPR consent. Its tracking script is also super tiny.
How it works
Yet another web analytics has three moving parts.
Your app or website sends analytics with a lightweight JavaScript library (yawa) that collects page views, custom events and Web Vitals.
On your self-hosted server, you deploy a Docker container that exposes an API and MCP server. It receives the events and stores them in an embedded DuckDB database, and provides 26+ analytics tools over the Model Context Protocol.
Lastly, connect your MCP client (Claude Code/Desktop/Web, Codex, or any MCP-compatible client) to your endpoint and query your data in natural language.
Quick Start
Run the server on any VPS or locally if you just want to give it a spin.
- Create a
docker-compose.yml
services:
app:
image: peterpeterparker/yawa:latest
ports:
- "3000:3000"
volumes:
- yawa-data:/data
environment:
- YAWA_SESSION_SECRET=${YAWA_SESSION_SECRET}
restart: unless-stopped
volumes:
yawa-data:
- Generate a secret
Optional, but improves visitor counting accuracy when generating the session IDs.
echo "YAWA_SESSION_SECRET=$(openssl rand -base64 32)" > .env
- Start the server
docker compose up -d
Create an access token
[!NOTE] Tokens are used to authenticate the MCP client.
docker exec -it <container-name> ./cli token create --name mytoken
Copy the token - it will only be shown once.
Register your site
[!NOTE] yawa supports multiple sites.
docker exec -it <container-name> ./cli site create --hostname yourdomain.com
Connect your MCP client
Once your server is running, connect any MCP-compatible client using the token created above.
Claude Code
claude mcp add yawa https://your-server.com/mcp --transport http \
--header "Authorization: Bearer YOUR_TOKEN"
Cursor
Add to your mcp.json:
{
"mcpServers": {
"yawa": {
"url": "https://your-server.com/mcp",
"headers": {
"Authorization": "Bearer YOUR_TOKEN"
}
}
}
}
Claude (web)
Go to Settings -> Integrations -> Add MCP server and enter your server URL and Bearer token.
Once connected, start by listing your sites:
List my analytics sites
Show me stats for yourdomain.com for the last 30 days
What are my top pages this week?
Install the tracker
Install the tracker in your website or frontend app:
npm install yawa-tracker
Setup
import { initYawa } from "yawa-tracker";
const cleanup = initYawa({
serverUrl: "https://your-yawa-server.com",
});
This automatically tracks page views on load and navigation (SPA-friendly via history.pushState and popstate).
Custom events
To track a custom event, call trackEvent with a name and optional metadata:
import { trackEvent } from "yawa-tracker";
trackEvent({ name: "signup", metadata: { plan: "pro" } });
Keys and values must be strings, with a maximum of 10 keys and 200 characters per key/value.
Web Vitals
Core Web Vitals (CLS, FCP, INP, LCP, TTFB) can also be collected by enabling the option when initializing the tracker:
import { initYawa } from "yawa-tracker";
const cleanup = initYawa({
serverUrl: "https://your-yawa-server.com",
webVitals: true,
});
API
| Function | Description |
|---|---|
initYawa(options) |
Initialize the tracker. Returns a cleanup function. |
trackPageView() |
Fire-and-forget page view tracking. |
trackPageViewAsync() |
Async page view tracking. |
trackEvent(data) |
Fire-and-forget custom event. |
trackEventAsync(data) |
Async custom event. |
Options
| Option | Required | Description |
|---|---|---|
serverUrl |
Yes | URL of your yawa server. |
webVitals |
No | Enable Web Vitals tracking. Defaults to false. |
MCP tools
Once connected, the following tools are available:
Sites
| Tool | Description |
|---|---|
list_sites |
List all registered sites with their IDs |
Page views
| Tool | Description |
|---|---|
get_stats |
Total pageviews, visitors, visits and bounces |
get_pageviews_series |
Daily pageviews and visitors over a date range |
get_pageviews_by_hour |
Pageviews by hour of day (0-23) |
get_top_pages |
Most visited URLs |
get_top_pages_expanded |
Top pages with engagement metrics |
get_top_titles |
Top page titles |
get_entry_pages |
Most common landing pages |
get_exit_pages |
Most common exit pages |
get_top_referrers |
Top referrers |
get_top_referrers_expanded |
Top referrers with engagement metrics |
get_browsers |
Browser breakdown |
get_operating_systems |
OS breakdown |
get_devices |
Device type breakdown |
get_languages |
Language breakdown |
get_time_zones |
Time zone breakdown |
get_utm_sources |
UTM source breakdown |
get_utm_mediums |
UTM medium breakdown |
get_utm_campaigns |
UTM campaign breakdown |
get_utm_contents |
UTM content breakdown |
get_utm_terms |
UTM term breakdown |
Custom events
| Tool | Description |
|---|---|
get_top_events |
Custom event names ranked by count |
get_event_series |
Daily event counts over a date range |
Web Vitals
| Tool | Description |
|---|---|
get_web_vitals_summary |
Average, p75 and p90 per metric (CLS, FCP, INP, LCP, TTFB) |
get_web_vitals_by_page |
Per-page breakdown for a specific metric |
get_web_vitals_distribution |
Good/needs improvement/poor counts per metric |
Configuration
| Variable | Required | Description |
|---|---|---|
YAWA_SESSION_SECRET |
Recommended | Secret used to hash session IDs. If not set, random UUIDs are used instead (less accurate visitor counting). |
YAWA_DATA_DIR |
No | Path where the DuckDB database file is stored. Defaults to /data. If modified, update the volume mount in your docker-compose.yml accordingly. |
Development
Clone the repo and run the app locally.
Setup
git clone https://github.com/peterpeterparker/yawa.git
cd yawa
bun install --frozen-lockfile
Run locally
bun run --filter yawa-app dev
The app server starts on http://localhost:3000 (events and MCP) and http://localhost:9999 (CLI only).
CLI
bun run --filter yawa-cli dev token create --name test
bun run --filter yawa-cli dev site create --hostname localhost
Run tests
bun test
Build
bun run --filter yawa-app build
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.