agentsvc.io Batch Web Page Reader

Use this endpoint when an AI agent needs to read and extract text from multiple web pages in a single round-trip, especially when injecting page content into an LLM context window. Prefer it over single-URL scrapers when you have 2-5 pages to fetch simultaneously, saving latency and cost. It is ideal when you need clean, readable Markdown rather than raw HTML, and when you want per-page metadata (title, status, word count) alongside content. Because it requires no API key and is paid per-call via USDC on Base, it is well-suited for autonomous agents that manage their own payment.

• URL is unreachable or DNS fails — page entry returned with ok:false and error message
• Page requires JavaScript rendering that the fetcher cannot execute — incomplete or empty markdown
• Target site blocks scrapers via robots.txt or rate limiting — fetch fails with error
• URL exceeds redirect limits or results in a redirect loop — fetch fails
• max_chars too low to capture meaningful content — markdown truncated with truncated:true
• Sending more than 5 URLs in one call — request rejected with validation error
• Invalid URL format in the array — request fails schema validation