# Live Google Search Console Data for Claude Code
<p align="center">
<a href="https://skills.sh/refreshagent/live-search-console-data"><img src="https://skills.sh/b/refreshagent/live-search-console-data" alt="Install via skills.sh"></a>
</p>
Skip the GCP OAuth setup. Give Claude Code live access to your Google Search Console and GA4 data in 60 seconds.
## Installation
```bash
npx skills add refreshagent/live-search-console-data
```
## How to Authenticate
After installing, ask Claude about your search traffic. The skill detects missing authentication, opens a secure RefreshAgent Google login page in your browser, saves the local key automatically, and then continues the request.
No manual `.env` editing is required in the normal flow.
## What You Can Ask
- "Which pages lost the most traffic last month?"
- "Show me my top keywords by impressions"
- "Is my sitemap indexing correctly?"
- "What's my organic traffic trend for the last 30 days?"
- "Do any of my pages compete for the same keyword?"
- "Which landing pages have the best conversion rate?"
## How It Works
This skill connects Claude Code to the RefreshAgent API, which acts as a secure proxy for Google Search Console and Google Analytics 4.
### Data Flow & Trust Model
1. **Claude starts the helper** when it needs live GSC or GA4 data.
2. **You authenticate with Google** via refreshagent.com's OAuth flow in your browser.
3. **RefreshAgent stores an encrypted OAuth token**: not your Google password, not your data. The token is scoped to the GSC/GA4 APIs you authorize.
4. **The helper saves a local RefreshAgent key** to `~/.config/refreshagent/.env` so future Claude Code sessions work without repeating setup.
5. **The AI agent calls RefreshAgent's API** using your API key (`ra_live_...`). RefreshAgent proxies the request to Google's official APIs using your stored token.
6. **Data is cached briefly** (seconds to minutes) to avoid redundant API calls. Cache age is reported in every response.
7. **No data leaves RefreshAgent for training.** Responses are ephemeral: they exist only in the agent's context window.
This is the same architecture used by every SEO tool that offers Google data integration — we handle the OAuth boilerplate so you don't have to configure a GCP project, set up a service account, or manage token refreshes.
### Why a Proxy Instead of Direct API Calls?
Google's Search Console and GA4 APIs require:
- A Google Cloud Project with OAuth 2.0 configured
- A registered redirect URI per client
- Handling token refresh flows
- Managing service account JSON files for server-to-server access
None of these work inside an AI agent's ephemeral execution environment. RefreshAgent handles this infrastructure so the agent can focus on answering your questions.
### Write Endpoints
This skill is primarily read-only, but does expose two write endpoints for creating client site mappings and proposals (`POST /api/v1/clients`, `POST /api/v1/proposals/build/*`). These require your explicit API key and only affect your own RefreshAgent account — they do not modify your Google Search Console or GA4 data.
## Requirements
- [Claude Code](https://claude.ai/download) (or any agent with skills.sh compatibility)
- A RefreshAgent account (free tier available)
- A Google account with Search Console or GA4 access
## Links
- [refreshagent.com](https://refreshagent.com)
- [skills.sh directory](https://skills.sh/refreshagent/live-search-console-data)
- [Agent Skills Format](https://agentskills.io)
- [Google Search Console API](https://developers.google.com/webmaster-tools)
- [Google Analytics Data API](https://developers.google.com/analytics/devguides/reporting/data/v1)