Ask Website with AI
Pricing
$5.00/month + usage
Ask Website with AI
Analyzes websites using AI (Gemini/OpenAI) to answer questions from scraped content. It can explore internal links for comprehensive answers, taking a list of URLs and questions. Ideal for targeted data extraction and content summarization.
5.0 (1)
Pricing
$5.00/month + usage
0
Total users
4
Monthly users
4
Runs succeeded
95%
Last modified
6 days ago
🧠 Ask Website with AI
The Ask Website with AI Apify Actor is designed to analyze website content and answer specific questions using advanced AI models. It combines web scraping with AI-powered natural language understanding to generate accurate answers—even exploring internal links when needed.
🚀 Features
- Apify SDK for Python — A robust toolkit for building Apify Actors and scrapers in Python.
- Input Schema — Defines and validates the structure of the Actor's input.
- Request Queue — Manages the list of URLs to be scraped efficiently.
- Dataset — Stores structured data with uniform attributes.
- Playwright — A browser automation library used for dynamic web scraping.
- AI Integration — Supports Google Gemini and OpenAI GPT models for analysis.
- Intelligent Link Exploration — Optionally explores relevant internal links if no answer is found on the initial page, using semantic relevance prioritization.
🧭 How It Works
-
Initial Page Scraping
Navigates to the given URL using Playwright and scrapes the HTML content, which is cleaned and structured into a JSON object. -
Initial AI Analysis
The AI model attempts to answer the provided questions using only the scraped content from the initial page. -
Internal Link Exploration (Optional)
If no clear answer is found andtopNLinksToExplore > 0, the Actor identifies and ranks internal links. The AI chooses the most relevant one(s) to explore further. -
Iterative Analysis
The scraping and AI analysis process repeats across selected links, up to the number defined intopNLinksToExplore. -
Combined Content Analysis
If no definitive answer is found from individual pages, the Actor combines all scraped content and requests a final answer from the AI model.
⚙️ Input Configuration
The Actor expects input in JSON format, validated by input_schema.json:
| Field | Type | Required | Description |
|---|---|---|---|
start_urls | array | ✅ Yes | A list of objects, each with a url string. Example:[{ "url": "https://apify.com" }] |
questions | array | ✅ Yes | List of questions for the AI model to answer. Example:["What is the topic of the site?", "What is the email of the site?"] |
instruction | string | ❌ No | Extra instructions for the AI, e.g., "return only the specific answer without any explanation or extra text" |
AIModel | string | ✅ Yes | Choose from:"gemini-1.5-flash" (Recommended)"gpt-4o-mini" (Recommended)"gemini-1.5-pro""gpt-4.1-mini""gpt-4.1" |
apiKey | string | ✅ Yes | Your API key for the selected AI model |
topNLinksToExplore | integer | ❌ No | Max internal links to explore if no answer is found. Default: 3. Set 0 to disable |
🛠 Getting Started
Build the Actor
Follow Apify’s documentation for instructions on building Actors.
Run the Actor
Provide the necessary start_urls, questions, AIModel, apiKey, and optional instruction and topNLinksToExplore.
🧪 Local Development
You can develop and test this Actor locally using the Apify CLI.
1. Install apify-cli
Using Homebrew:
