RAG Web Search: Search Results as Markdown for LLMs
Web search for AI agents, LLMs and RAG: run any query and get the top result pages as clean Markdown with title, snippet, author, dates and word count. Fast HTTP, no browser, ideal for tool calls and MCP. Pay per result.
Overview
A web-search Actor that performs live web queries and returns top search results formatted as clean, LLM-ready Markdown for retrieval-augmented generation (RAG), agent tooling and citation workflows. For each result it provides search rank, title, URL, snippet, and—when page fetching is enabled—the page's main content extracted to Markdown using a readability-style extractor (menus, ads and banners removed) plus metadata such as description, author, published/modified dates, language and word count. The Actor is HTTP-based (no browser), optimized for fast bulk retrieval, deduplicates results, substitutes failed or blocked pages with the next results, and supports SERP-only mode, character-limited page content to fit context windows, domain skipping, region and time-range filtering. Search sources combine DuckDuckGo results with supplemental search results when larger result sets are requested; outputs are structured for grounding, citation, summarization and RAG pipelines.
Use cases
- Provide live web access for LLMs and AI agents (ChatGPT, Claude, LangChain, LlamaIndex and similar).
- Grounding and fact-checking model answers with fresh web sources and citations.
- Research automation and topical monitoring using time-range filtering to collect recent coverage.
- Content and SEO research to inspect what top-ranking pages actually say and extract article content for analysis.
Pricing
| Result scraped | $0.003 per event (3 USD per 1,000) |
|---|---|
| Billing | Charged by Apify to your Apify account; platform usage may apply |
How to use
Open on Apify Store
Sign in to Apify or create a free account.
Configure the input
Fill in the input form; each field is documented on the Apify page.
Run and export
Start the run and download results as JSON, CSV or Excel, or use the API.
