Skip to main content

Changelog · September 25, 2026

MCP: prompts, a pricing resource, structured results, and two renamed tools

Ready-made workflows for your agent, and results a client can parse without guessing.

By · Last updated: September 2026

TL;DR

The MCP server now offers three prompts (read_page, audit_page, compare_pages), the plans and credit prices as the resource urlpipe://pricing, and an output schema with structured results on all 14 tools. Two tools were renamed so every name starts with what it does: console_logs is capture_console_errors and lighthouse_audit is run_lighthouse_audit. The server lists its tools, prompts and resources without a token; calling anything still needs one.

Free plan, no credit card. 1,000 credits a month.

Renamed tools. console_logs is now capture_console_errors and lighthouse_audit is now run_lighthouse_audit. Arguments and results are unchanged; update any allow-list or prompt that names them.

Prompts. read_page answers a question from a page, audit_page turns Lighthouse, console errors and a screenshot into a short health report, and compare_pages puts two URLs side by side. In Claude Code they appear as /mcp__urlpipe__… commands. Each tells the agent which tools to call and what they cost; the calls are billed as usual.

Structured results. Every tool declares an output schema and returns its result as structuredContent as well as text: the object itself for metadata, Lighthouse and scrapes, and {"markdown": …}, {"keywords": […]} and the like where the result isn't an object. A screenshot stays one image; its structured result is its type and link.

FAQ

Frequently asked questions

No. capture_console_errors and run_lighthouse_audit take the same arguments and return the same results as before; only the names changed.

Try it on the free plan.

Free plan, no card. Confirm your email and your API key is live — you'll be making real requests in minutes.