🌐 RAG Web Browser (byseitz.agency/rag-browser) This Actor provides essential web browsing and content extraction functionality for AI Agents, LLM applications, and Retrieval-Augmented Generation (RAG) pipelines. It functions similarly to the web search feature in popular LLM chatbots, providing fresh, contextualized data directly from the web. This Actor is published using the Pay-per-Event (PPE) monetization model, ensuring predictable pricing where you are charged only for the precise actions performed. --- ## Main features and Purpose * Real-Time Grounding: Accepts a search phrase or URL and queries Google Search to get up-to-date information. * Structured Output: Cleans HTML from web pages and returns content as clean Markdown, making it immediately ready for LLM processing. * Hybrid Scraping for Speed: Uses a fast, non-browser approach (raw-http) to crawl result pages quickly. When anti-bot systems block full content, the Actor reliably returns search result descriptions and essential metadata. * AI/LLM Compatibility: Fully compatible with popular LLM frameworks (LangChain, CrewAI, etc.) for direct tool integration. --- ## 💰 Pay-per-Event (PPE) Pricing Breakdown Under the PPE model, you do not pay for underlying Compute Units (CUs) or typical platform usage for the Actor run. Instead, you pay a fixed fee for each valuable event triggered. | Event Name | Unit Charged | Price | Description | | :--- | :--- | :--- | :--- | | apify-actor-start | Per run | $0.00005 | The compulsory charge for initiating the Actor run. | | RAG_QUERY | Per query | $0.002 | Charges for executing one search query, including the Google SERP lookup and the attempt to crawl up to 3 target web pages. | ### Note on Cost The price for the RAG_QUERY event is set to cover the consumption costs (proxy, SERP lookups) generated by this specialized tool and maintain its performance. Your final invoice will only reflect the sum of these charged events and low, standard charges for data storage. --- ## ⚙️ Usage and Integration ### Input The Actor accepts the following parameters (among others): | Parameter | Default | Description | | :--- | :--- | :--- | | query | (Required) | The Google search term or a single full URL to crawl. | | maxResults | 3 | The maximum number of top search results to crawl from the SERP. | | scrapingTool | raw-http | Recommended: Use raw-http for speed and cost efficiency. Use browser-playwright only for known JavaScript-heavy sites. | | requestTimeoutSecs | 45 | Maximum time (in seconds) allowed for search and content extraction per request. | ### Output Example The output is an array of JSON objects, typically including: | Field | Description | | :--- | :--- | | metadata.url | The URL of the scraped page. | | searchResult.description | The original Google search snippet (used as fallback content). | | markdown | The clean, extracted text content in Markdown format. |