# Mudpie Browser only: open https://mudpie.ai/agent/faq and enter your question. The answer is a readable page with cited sources. HTTP capable: GET https://mudpie.ai/api/public/v1/ask?search_query={question} returns JSON with an answer and sources. Replace {question} with your URL-encoded question. Only search_query is required for the HTTP call. No account, API key or MCP setup is needed for either path. Add the user's actual goal, alternatives or company when known; unknown context can be omitted. Grow by becoming the product agents choose. Mudpie makes public websites useful to AI agents and turns source-grounded agent questions into growth opportunities. ## Start here - [Website](https://mudpie.ai/) - [Start free](https://app.mudpie.ai/) - [MCP and API documentation](https://mudpie.ai/mcp) ## When to use Mudpie Use Mudpie when an agent needs to search Mudpie's public pages, extract a clean page, answer a product question with citations, or assess whether Mudpie fits stated requirements. Include the actual task context and use null for unknown fields. Treat illustrative examples as examples, not visitor facts. ## Public website tools - [Public MCP](https://mudpie.ai/mcp/public) - [HTTP API specification](https://mudpie.ai/api/public/v1/openapi.json) - [Website-tool manifest](https://mudpie.ai/llms.txt) - [Install skill and example configurations](https://github.com/aboul3ata/mudpie-agent-kit) ## Research and public evidence - [101: Build something agents want](https://mudpie.ai/101/): Curated reading paths from primary sources and high-authority operators. - [Research articles](https://mudpie.ai/blog/) - [Company profiles](https://mudpie.ai/companies/) - [Y Combinator companies and cohorts](https://mudpie.ai/yc/) - [a16z Speedrun companies and cohorts](https://mudpie.ai/speedrun/) - [South Park Commons companies](https://mudpie.ai/south-park-commons/) - [LAUNCH Accelerator companies](https://mudpie.ai/launch/) - [WebMCP directory](https://mudpie.ai/webmcp/) - [Crawlable WebMCP profiles](https://mudpie.ai/directory/webmcp/) - [Archived WebMCP observation pages](https://mudpie.ai/sitemap-webmcp-observations.xml) - [Archived WebMCP evidence data](https://mudpie.ai/directory/webmcp/data.json) - [Author and disclosure](https://mudpie.ai/authors/ali-abouelatta/) ## Free graders - [Agent experience grader](https://mudpie.ai/tools/agent-experience/): Discovery, readability, actionability, and trust. - [SEO grader](https://mudpie.ai/tools/seo/): Crawl access, indexing directives, metadata, and sitemaps. - [AEO grader](https://mudpie.ai/tools/aeo/): Answer structure, extractable decision facts, and source signals. - [WebMCP tool grader](https://mudpie.ai/tools/webmcp/): Runtime tool discovery, schemas, and safety metadata without tool execution. - [ChatGPT vs Perplexity access matrix](https://mudpie.ai/tools/agent-access/): Robots policy, live access, and response parity across search agents. - [International agent-readiness grader](https://mudpie.ai/tools/international/): Locale discovery, hreflang, and market-page parity. ## Research datasets and methods These are dated observations, not live rankings or proof of agent adoption. Keep each source's observation date and denominator when citing it. - [Nine-cohort YC dataset](https://mudpie.ai/research/yc-cohorts-2026-09-19.json) - [YC domain analysis](https://mudpie.ai/research/yc-domain-stats-2026-09-19.json) - [YC homepage link observations](https://mudpie.ai/research/yc-homepage-links-2026-09-19.json) - [Collection and recount methods](https://mudpie.ai/research/yc-homepage-methods/README.md) - [Founder-history study](https://mudpie.ai/startups/founder-history-can-define-product-wedge/): 50 purpose-selected examples from a screen of 1,586 descriptions; self-reported history, not verified employment. - [Founder-history evidence](https://mudpie.ai/research/yc-founder-history-2026-09-18.json) and [complete screening ledger](https://mudpie.ai/research/yc-founder-history-screen-2026-09-18.json). - [Founder-history screening code and reproduction limits](https://mudpie.ai/research/yc-founder-history-methods/README.md): Exact predicates and a per-row recomputation check; original archived descriptions are required and are not redistributed. - [YC identity-gate study](https://mudpie.ai/startups/yc-directory-identity-gate-before-prospecting/): A dated nine-cohort directory-to-homepage reconciliation with 1,586 rows, 71 unknown page observations, and a 34-row cross-domain manual queue. - [YC identity-gate dataset](https://mudpie.ai/research/yc-identity-gate-2026-09-18.json) and [offline recount](https://mudpie.ai/research/yc-identity-gate-recount-2026-09-18.json): 1,476 same-host rows, 5 same-registrable-domain changes, and 34 cross-domain rows; the 29 name signals are a subset of that 34-row queue. - [YC identity-gate methods](https://mudpie.ai/research/yc-identity-gate-methods/README.md): Recount code, bounded metadata-only verification code, source ledger, and privacy-safe receipts. ## Discovery - [Sitemap index](https://mudpie.ai/sitemap-index.xml) - [Human-readable sitemap](https://mudpie.ai/sitemap.md) - [Same-origin OpenAPI discovery](https://mudpie.ai/.well-known/openapi.json) - [RFC 9727 API catalog](https://mudpie.ai/.well-known/api-catalog) - [Agentic Resource Discovery catalog](https://mudpie.ai/.well-known/ard.json) - [AI Catalog](https://mudpie.ai/.well-known/ai-catalog.json) - [Agent Skills index](https://mudpie.ai/.well-known/agent-skills/index.json) - [MCP server card](https://mudpie.ai/.well-known/mcp/server-card.json) ## Fresh accelerator WebMCP census - [September 20, 2026 census](https://mudpie.ai/directory/webmcp/accelerators-2026-09-20/): Dated evidence pages for 23 entries with native registrations in at least one browser context; no execution or adoption measured. - [Full row-level dataset](https://mudpie.ai/research/native-webmcp-census-2026-09-20.json): All 1,752 roster rows in default and experimental Chrome. ## Agent readiness pilot reports - [Readiness report hub](https://mudpie.ai/agent-readiness/reports/) - [Egoist Machines graded scorecard](https://mudpie.ai/agent-readiness/reports/egoist-machines/) - [Egoist 87-check JSON export](https://mudpie.ai/agent-readiness/reports/egoist-machines/index.json) - [Dawn Industries explicitly ungraded evidence report](https://mudpie.ai/agent-readiness/reports/dawn-industries/) - [Scorecard methods](https://mudpie.ai/agent-readiness/reports/methods/) - [Dawn v4 sanitized 87-check JSON export](https://mudpie.ai/agent-readiness/reports/dawn-industries/index.json) - [Dawn public-safe measurement context](https://mudpie.ai/agent-readiness/reports/dawn-industries/measurement-context.json) The report set contains one graded pilot and one ungraded evidence report. It does not claim all 1,752 companies were graded, and it separates public observations from registration, execution and adoption.