page-speed-insights-mcp
An MCP server that wraps the Google PageSpeed Insights API to analyse web performance, providing tools for scores, Core Web Vitals, opportunities, diagnostics, and batch analysis.
README
PageSpeed Insights MCP Server
An MCP server that wraps the Google PageSpeed Insights API v5, exposing 7 focused tools for analysing web performance — scores, Core Web Vitals, opportunities, diagnostics, and batch analysis.
PWA Category Notice: The
pwaLighthouse category was removed in Lighthouse 12+ and is no longer available via the PSI API. Requesting it returns an API error.psi_check_pwais intentionally not implemented.
Features
| Tool | Description |
|---|---|
psi_analyze_url |
Full analysis — scores + CWV + top-5 opportunities |
psi_get_core_web_vitals |
Lightweight CrUX field data + lab CWV |
psi_get_opportunities |
Opportunities sorted by estimated savings |
psi_get_diagnostics |
Diagnostic issues (DOM size, JS work, etc.) |
psi_compare_mobile_desktop |
Side-by-side mobile vs desktop comparison |
psi_batch_analyze |
Analyse up to 10 URLs, partial-failure tolerant |
psi_get_category_score |
Single category score + audit breakdown |
Prerequisites
- Python 3.11+
- A Google API key with the PageSpeed Insights API enabled
Step 1 — Get a Google API Key
- Go to the Google Cloud Console.
- Create or select a project.
- Navigate to APIs & Services → Library and search for "PageSpeed Insights API".
- Click Enable.
- Go to APIs & Services → Credentials → Create Credentials → API key.
- Copy the generated key.
Quota: 25,000 requests/day · 400 requests/100 seconds per API key.
To request higher limits: Cloud Console → IAM & Admin → Quotas.
Step 2 — Install
With pip (virtualenv recommended)
cd page-speed-insights-mcp
python -m venv .venv
# Windows
.venv\Scripts\activate
# macOS/Linux
source .venv/bin/activate
pip install -r requirements.txt
With uv (recommended)
cd page-speed-insights-mcp
uv venv
uv pip install -r requirements.txt
Step 3 — Set PSI_API_KEY
Copy the example file and add your key:
cp .env.example .env
Edit .env:
PSI_API_KEY=AIzaSy...your_key_here
The server loads this automatically via python-dotenv. Alternatively, export it directly:
# Windows PowerShell
$env:PSI_API_KEY = "AIzaSy..."
# macOS/Linux
export PSI_API_KEY="AIzaSy..."
Security: Never commit your
.envfile or hardcode your API key in source code.
Step 4 — Verify the Installation
# Syntax check all modules
python -m py_compile pagespeed_mcp/server.py
python -m py_compile pagespeed_mcp/api.py
python -m py_compile pagespeed_mcp/models.py
python -m py_compile pagespeed_mcp/formatters.py
echo "All files compiled OK"
Step 5 — Register with Claude Desktop
Edit your Claude Desktop config file:
- macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Windows:
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"pagespeed": {
"command": "python",
"args": ["-m", "pagespeed_mcp.server"],
"cwd": "D:/SIVA/mcps/page-speed-insights-mcp",
"env": {
"PSI_API_KEY": "AIzaSy...your_key_here"
}
}
}
}
Tip: Use the full path to the Python executable inside your virtualenv instead of bare
pythonto avoid PATH issues:
- Windows:
"D:/SIVA/mcps/page-speed-insights-mcp/.venv/Scripts/python.exe"- macOS/Linux:
"/path/to/page-speed-insights-mcp/.venv/bin/python"
Step 6 — Register with Claude Code
In your project's .claude/settings.json or ~/.claude/settings.json:
{
"mcpServers": {
"pagespeed": {
"command": "python",
"args": ["-m", "pagespeed_mcp.server"],
"cwd": "D:/SIVA/mcps/page-speed-insights-mcp",
"env": {
"PSI_API_KEY": "AIzaSy...your_key_here"
}
}
}
}
Or add via CLI:
claude mcp add pagespeed python -m pagespeed_mcp.server \
--cwd "D:/SIVA/mcps/page-speed-insights-mcp" \
--env PSI_API_KEY=AIzaSy...
Step 7 — Test with MCP Inspector
npx @modelcontextprotocol/inspector python -m pagespeed_mcp.server
This opens a browser UI where you can call each tool interactively.
Example Tool Calls
1. Full analysis (Markdown)
{
"tool": "psi_analyze_url",
"arguments": {
"url": "https://web.dev",
"strategy": "mobile",
"categories": ["performance", "accessibility", "seo"]
}
}
2. Quick Core Web Vitals check
{
"tool": "psi_get_core_web_vitals",
"arguments": {
"url": "https://www.google.com",
"strategy": "desktop"
}
}
3. Opportunities with minimum savings
{
"tool": "psi_get_opportunities",
"arguments": {
"url": "https://example.com",
"strategy": "mobile",
"min_savings_ms": 200
}
}
4. Diagnostics
{
"tool": "psi_get_diagnostics",
"arguments": {
"url": "https://example.com",
"strategy": "mobile"
}
}
5. Mobile vs Desktop comparison (JSON output)
{
"tool": "psi_compare_mobile_desktop",
"arguments": {
"url": "https://web.dev",
"response_format": "json"
}
}
6. Batch analysis
{
"tool": "psi_batch_analyze",
"arguments": {
"urls": [
"https://web.dev",
"https://www.google.com",
"https://github.com"
],
"strategy": "mobile",
"categories": ["performance", "seo"]
}
}
7. Single category score
{
"tool": "psi_get_category_score",
"arguments": {
"url": "https://example.com",
"category": "accessibility",
"strategy": "desktop"
}
}
Caching
Repeated calls for the same (url, strategy, categories) within a 5-minute window return cached results, avoiding unnecessary API quota consumption. The cache is in-memory and resets when the server restarts.
Rate Limits & Best Practices
| Limit | Value |
|---|---|
| Requests per day | 25,000 |
| Requests per 100 seconds | 400 |
- Use
psi_get_core_web_vitalsfor quick health checks (only fetchesperformancecategory). - In batch mode, specify only the categories you need to reduce response time.
- The 5-minute cache prevents redundant calls in a single session.
- On HTTP 429, the server automatically retries up to 3 times with back-off.
Project Structure
page-speed-insights-mcp/
├── pagespeed_mcp/
│ ├── __init__.py # Package marker
│ ├── server.py # FastMCP app + 7 tool definitions
│ ├── api.py # HTTP client, cache, retry, error handling
│ ├── models.py # Pydantic v2 input models
│ └── formatters.py # Response extractors and Markdown/JSON renderers
├── requirements.txt
├── .env.example
└── README.md
Troubleshooting
| Error | Fix |
|---|---|
❌ Missing API key |
Set PSI_API_KEY in .env or environment |
❌ API key rejected (HTTP 403) |
Verify the key is valid and PageSpeed Insights API is enabled in your Cloud project |
❌ Invalid URL (HTTP 400) |
Ensure the URL is publicly accessible and starts with https:// |
⏳ Rate limited (HTTP 429) |
Reduce request frequency; the server retries automatically up to 3 times |
⏱️ Request timed out |
The page may be unreachable or very slow; try again |
pwa category error |
PWA category was removed from Lighthouse 12+ — do not pass pwa |
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.
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.
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.
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
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.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.