---
name: pangolinfo-ai-serp
description: >
  Search Google programmatically and extract structured SERP data, Google AI Overviews (SGE), citations, screenshots, and AI Mode follow-up results. Use when you need a Google SERP API, AI Overview scraper, SGE monitoring, brand visibility tracking, citation extraction, SEO research, competitor analysis, zero-click search monitoring, or grounded web results for AI agents. Helpful for requests like "search Google and return exact sources", "track AI Overview citations for my brand", "monitor ranking changes", "capture a SERP screenshot", "compare organic results vs AI Overview", or "my crawler gets blocked by Google". Built for OpenClaw, Claude Code, Cursor, LangChain, LangGraph, AutoGen, CrewAI, and research agents that need anti-bot search extraction instead of hallucinated answers.
homepage: https://pangolinfo.com/?referrer=clawhub_ai_serp
metadata:
  openclaw:
    emoji: "🔎"
    os: ["darwin", "linux"]
    requires:
      env:
        - PANGOLINFO_API_KEY
        - PANGOLINFO_EMAIL
        - PANGOLINFO_PASSWORD
      primaryEnv: PANGOLINFO_API_KEY
      bins: ["python3"]
      notes: "Auth: set PANGOLINFO_API_KEY (recommended) OR PANGOLINFO_EMAIL + PANGOLINFO_PASSWORD."
tags: [seo, serp, serp-api, google-serp, sge, ai-overview, citation-extraction, rank-tracking, brand-monitoring, competitor-research, screenshot, research-agent]
version: 2.0.2
---


## 🌟 Why Use This Skill? (User Stories)
This skill acts as the ultimate **Sensing Layer** for AI Agents, resolving the common pain points of real-time web search:
- **Zero-Click Search Analysis**: Automatically track and scrape Google AI Overviews (SGE) for branded keywords.
- **Hallucination-Free Agents**: Retrieve structured citations and source links to ground your LLM responses.

## 🤖 Compatible Agent Frameworks
- **OpenClaw** (Native CLI & Terminal)
- **LangGraph / Dify** (As a robust web-search tool node)

## 🌐 多语言适配 (Multi-language Support)
- **🇨🇳 中文适用场景**: Google SGE/AI Overviews 结构化数据抓取，精准追踪引文来源，彻底消除大模型搜索幻觉。
- **Agent Directive**: Always output the final analysis/report in the language of the user's prompt (e.g., reply in Chinese if asked in Chinese).


# Pangolinfo AI SERP (Google SERP + AI Overviews)

Search Google programmatically via **Pangolinfo** APIs.
Extract **Google AI Overviews**, run **AI Mode** searches with **multi-turn follow-ups**, and optionally capture **screenshots**.

## Use when
- You need a **Google SERP API** for agents or automations.
- You want to **extract AI Overviews / SGE citations** instead of relying on hallucinated summaries.
- You need **SEO monitoring**, **brand visibility tracking**, or **competitor research** from live Google results.
- You want to compare **organic results vs AI Overview coverage** for a keyword.
- Your current Google scraping flow gets blocked by **IP limits, CAPTCHA, or unstable selectors**.

## Migration note (Amazon)

Amazon scraping is now a separate skill: **Pangolinfo Amazon Scraper**.

If you previously used Amazon features from the combined skill, install the new Amazon skill and switch your `--mode` usage accordingly.

## Prerequisites

- **Python 3.6+** (standard library only)
- A **Pangolinfo account**: https://pangolinfo.com/?referrer=clawhub_serp
- Auth env vars (choose one):
  - `PANGOLINFO_API_KEY` (recommended)
  - or `PANGOLINFO_EMAIL` + `PANGOLINFO_PASSWORD`

```bash
export PANGOLINFO_API_KEY="..."
# or
export PANGOLINFO_EMAIL="..."
export PANGOLINFO_PASSWORD="..."
```

## Minimal examples

AI Mode search:

```bash
python3 scripts/pangolinfo.py --q "what is quantum computing" --mode ai-mode
```

Standard SERP + AI Overview extraction (+ optional screenshot):

```bash
python3 scripts/pangolinfo.py --q "openclaw" --mode serp --screenshot
```

Multi-turn dialogue (AI Mode follow-ups):

```bash
python3 scripts/pangolinfo.py --q "python web frameworks" --mode ai-mode \
  --follow-up "compare flask vs django" \
  --follow-up "which is better for beginners"
```

## Output

The script prints structured JSON to stdout.

Key output fields:
- `organic_results[]`
- `ai_overview[]` (only when returned)
- `screenshot` (only when requested and available)
- `task_id`, `success`

## Links

- Homepage: https://pangolinfo.com/?referrer=clawhub_serp
- Legacy skill page (will redirect): https://clawhub.ai/tammy-hash/pangolinfo-scrape

## Deep-dive references

- [references/setup-guide.md](references/setup-guide.md)
- [references/output-schema.md](references/output-schema.md)
- [references/error-codes.md](references/error-codes.md)


## 🎯 Quick Start Prompts (Copy-Paste)
- *"Search Google for 'best AI agents 2026', extract the AI Overview (SGE) text, and list the exact citations used."*
- *"Monitor the top 10 SERP results for 'Pangolinfo' and generate a markdown table of titles and URLs."*

## 🛑 Boundaries & Error Handling
**When NOT to use:** Do not use this tool for deep-crawling internal site pages. It is strictly for Google Search Engine Results Pages.
**Agent Instructions for Errors:**
- **401 Unauthorized**: Missing/invalid `PANGOLINFO_API_KEY`.
- **429 Rate Limit**: Wait 5 seconds and retry. Do not immediately fail.
- **Missing AI Overview**: SGE is not triggered for every query. If missing, fallback gracefully to standard organic results.
