Enables AI agents to perform company due diligence, OSINT, competitive, SEO, market, finance and regulatory research through a single MCP endpoint exposing 45 tools that draw on official public APIs, local D1 mirrors, and optional self-hosted sidecars. Every response is labelled by evidence class, so inferred estimates are never presented as equivalent to official data.
An MCP server that lets agents collect a company's entire public marketing footprint from a single URL — paid ads, organic social, SEO, AI-crawler stance, Reddit and founder activity, tracking pixels, and funnel signals — and return a compact scorecard plus an evidence index with unknowns. It also exposes tools to pull the underlying ad copy, posts and media, retrieve the analysis playbook, diff two runs over time, and validate that a generated report's citations resolve without invented spend.
Enables AI agents to call 16 small, deterministic, read-only tools over a free, no-auth remote streamable HTTP MCP endpoint, covering business idea scoring, price headroom, hotel OTA commissions, open-source SaaS swaps, ads policy notice risk, App Store 4.2 prechecks, and Japan trip planning. Each tool is also exposed as a plain JSON API with OpenAPI docs.
Machine-readable AI market data for AI agents: subscription prices, free credits, free LLM models, China-exclusive AI pricing. 6 MCP tools, Streamable HTTP, no auth.
Enables AI clients to discover, inspect, and run over 2,000 AI models and APIs through local MCP tools, with free data API calls and browser-based authorization. It supports workflows such as search, social media, scraping, e-commerce, finance, image/video/audio generation, and run/account management.