ResearchPi: Get an evidence-backed company profile from its official domain, reading at most 3 pages. Uses matching Organization data or a site brand corroborated by its page title; these are distinguished from verified legal identity. Returns cited name and description, links, missing fields and conflicts. Does not estimate revenue, funding or headcount.

ResearchPi: Extract readable text, title and links from one public HTTPS HTML page. Source URL and retrieval time included. No JavaScript, PDF, login or private network access. Maximum 2 MB HTML / 60,000 text characters.

ResearchPi: Check current SaaS headline plan prices, listed features and observed changes for Linear, Supabase, Firecrawl or Clerk. Source URL, retrieval time, missing currency/terms and annual billing are explicit. Optional compare_to snapshot ID; first observation creates a baseline. No complete historical coverage, usage-cost estimate or checkout quote. Failed extraction is not settled. Supported products: GET /pricing-products.

TokenBrief: Analyze a Base ERC-20 address with block-linked supply, owner/proxy observations and Uniswap v2/v3 + Aerodrome Slipstream pools (USDC/WETH). Returns spot quotes, pool inventory and evidence; wallet checkout and PNG cards at /tokenbrief. Optional narration=ai adds OpenAI commentary; default rules. No security verdict, holder concentration or sellability test. Cache up to 60s, disclosed. Source or AI failures are not settled.