Configure screenshot options, choose device presets, and copy ready-to-use cURL, JavaScript, or Python requests. Build requests without an account, or add your API key to render a screenshot.
New GET /extract endpoint and extract_text MCP tool: agents asking 'what does this page say' now get markdown/text instead of a 500 KB PNG. 5-10x faster, 100x cheaper LLM context.
Per-user webhook signing secrets stored with AES-256-GCM encryption; MCP server now requires OAuth 2.0 / OIDC so AI agents authenticate with their own credentials; webhook delivery SSRF protection and configurable timeout; encrypted JWT storage in the dashboard.