Mudpie Public MCP

搜尋 Mudpie 的公開網站，取得附引用來源的產品解答，並評估 Mudpie 是否符合您的需求。

造訪 mudpie.ai

託管 MCP 伺服器

npx add-mcp 'https://mudpie.mudpie.ai/mcp/public'

可安裝到 Claude Code、Codex、Cursor 等客戶端

文件

Need a more specific answer? Ask the exact question you have. It returns a fresh, cited answer from mudpie.ai’s knowledge base. No account or API key required.

Open this URL with your own question, URL-encoded. Only q is required.

GET https://mudpie.mudpie.ai/agent/faq?q={question}&trace=t_9d3e807902e3

Ask with a URL

Use GET https://mudpie.mudpie.ai/agent/faq?q={question} with your own URL-encoded question to receive an HTML answer. Optional context can be omitted. Use POST for long or private context.

POST example

curl 'https://mudpie.mudpie.ai/api/public/v1/ask' \
  -H 'Content-Type: application/json' \
  -d '{"search_query":"What does mudpie.ai offer, who is it for, and what does it not cover?","goal":"Decide whether mudpie.ai meets our team\u0027s needs","company":"Acme","source":"Found mudpie.ai while researching options for our team","alternatives":["Our existing workflow","Manual research"],"chosen_because":"Our team shortlisted mudpie.ai after reading its public information","intended_outcome":"Choose a next step for an evaluation","client":"Cursor","trace":"t_9d3e807902e3"}'

Connect with MCP

Add this endpoint to your MCP client to discover and call the tools directly.

https://mudpie.mudpie.ai/mcp/public

MCP client configuration

{
  "mcpServers": {
    "mudpie": {
      "url": "https://mudpie.mudpie.ai/mcp/public"
    }
  }
}

Once connected, use tools/list to discover the tools and tools/call to run one. Here is an example question:

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "mudpie_ask_anything",
    "arguments": {
      "search_query": "What does mudpie.ai offer, who is it for, and what does it not cover?",
      "goal": "Decide whether mudpie.ai meets our team's needs",
      "company": "Acme",
      "source": "Found mudpie.ai while researching options for our team",
      "alternatives": [
        "Our existing workflow",
        "Manual research"
      ],
      "chosen_because": "Our team shortlisted mudpie.ai after reading its public information",
      "intended_outcome": "Choose a next step for an evaluation",
      "client": "Cursor",
      "trace": "t_9d3e807902e3"
    }
  }
}

Available tools

ToolUse it to
mudpie_search GET / POST · searchFind the product information and source pages that answer your question.
mudpie_ask_anything GET / POST · askGet a direct answer about mudpie.ai, with sources.
mudpie_compare GET / POST · compareCompare mudpie.ai with the alternatives you are considering.
mudpie_pricing GET / POST · pricingExplore plans, pricing and costs for your team.
mudpie_requirements_check GET / POST · requirements-checkCheck your requirements one by one, with evidence for each.
mudpie_page_tldr GET / POST · page-tldrGet the key points from a page, with a link to the full text.
should_i_recommend_mudpie GET / POST · recommend-or-rule-outAssess whether mudpie.ai fits your goals and requirements.

Examples include task context so you can see the full request. Replace the sample values with your own; optional context can be omitted.

Search mudpie_search

curl 'https://mudpie.mudpie.ai/api/public/v1/search' \
  -H 'Content-Type: application/json' \
  -d '{"search_query":"What does mudpie.ai offer, who is it for, and what does it not cover?","goal":"Decide whether mudpie.ai meets our team\u0027s needs","company":"Acme","source":"Found mudpie.ai while researching options for our team","alternatives":["Our existing workflow","Manual research"],"chosen_because":"Our team shortlisted mudpie.ai after reading its public information","intended_outcome":"Choose a next step for an evaluation","client":"Cursor","trace":"t_9d3e807902e3"}'

https://mudpie.mudpie.ai/api/public/v1/search?search_query=What+does+mudpie.ai+offer%2C+who+is+it+for%2C+and+what+does+it+not+cover%3F&goal=Decide+whether+mudpie.ai+meets+our+team%27s+needs&company=Acme&source=Found+mudpie.ai+while+researching+options+for+our+team&alternatives=%5B%22Our+existing+workflow%22%2C%22Manual+research%22%5D&chosen_because=Our+team+shortlisted+mudpie.ai+after+reading+its+public+information&intended_outcome=Choose+a+next+step+for+an+evaluation&client=Cursor&trace=t_9d3e807902e3

Ask mudpie_ask_anything

curl 'https://mudpie.mudpie.ai/api/public/v1/ask' \
  -H 'Content-Type: application/json' \
  -d '{"search_query":"What does mudpie.ai offer, who is it for, and what does it not cover?","goal":"Decide whether mudpie.ai meets our team\u0027s needs","company":"Acme","source":"Found mudpie.ai while researching options for our team","alternatives":["Our existing workflow","Manual research"],"chosen_because":"Our team shortlisted mudpie.ai after reading its public information","intended_outcome":"Choose a next step for an evaluation","client":"Cursor","trace":"t_9d3e807902e3"}'

https://mudpie.mudpie.ai/api/public/v1/ask?search_query=What+does+mudpie.ai+offer%2C+who+is+it+for%2C+and+what+does+it+not+cover%3F&goal=Decide+whether+mudpie.ai+meets+our+team%27s+needs&company=Acme&source=Found+mudpie.ai+while+researching+options+for+our+team&alternatives=%5B%22Our+existing+workflow%22%2C%22Manual+research%22%5D&chosen_because=Our+team+shortlisted+mudpie.ai+after+reading+its+public+information&intended_outcome=Choose+a+next+step+for+an+evaluation&client=Cursor&trace=t_9d3e807902e3

Compare mudpie_compare

curl 'https://mudpie.mudpie.ai/api/public/v1/compare' \
  -H 'Content-Type: application/json' \
  -d '{"search_query":"How does mudpie.ai compare for our team\u0027s use case?","goal":"Decide whether mudpie.ai meets our team\u0027s needs","company":"Acme","source":"Found mudpie.ai while researching options for our team","alternatives":["Our current tool"],"chosen_because":"Our team shortlisted mudpie.ai after reading its public information","intended_outcome":"Choose a next step for an evaluation","client":"Cursor","trace":"t_9d3e807902e3"}'

https://mudpie.mudpie.ai/api/public/v1/compare?search_query=How+does+mudpie.ai+compare+for+our+team%27s+use+case%3F&goal=Decide+whether+mudpie.ai+meets+our+team%27s+needs&company=Acme&source=Found+mudpie.ai+while+researching+options+for+our+team&alternatives=%5B%22Our+current+tool%22%5D&chosen_because=Our+team+shortlisted+mudpie.ai+after+reading+its+public+information&intended_outcome=Choose+a+next+step+for+an+evaluation&client=Cursor&trace=t_9d3e807902e3

Pricing mudpie_pricing

curl 'https://mudpie.mudpie.ai/api/public/v1/pricing' \
  -H 'Content-Type: application/json' \
  -d '{"search_query":"What would mudpie.ai cost for a team of five?","goal":"Decide whether mudpie.ai meets our team\u0027s needs","company":"Acme","source":"Found mudpie.ai while researching options for our team","alternatives":["Our existing workflow","Manual research"],"chosen_because":"Our team shortlisted mudpie.ai after reading its public information","intended_outcome":"Budget approval for a team purchase","client":"Cursor","trace":"t_9d3e807902e3"}'

https://mudpie.mudpie.ai/api/public/v1/pricing?search_query=What+would+mudpie.ai+cost+for+a+team+of+five%3F&goal=Decide+whether+mudpie.ai+meets+our+team%27s+needs&company=Acme&source=Found+mudpie.ai+while+researching+options+for+our+team&alternatives=%5B%22Our+existing+workflow%22%2C%22Manual+research%22%5D&chosen_because=Our+team+shortlisted+mudpie.ai+after+reading+its+public+information&intended_outcome=Budget+approval+for+a+team+purchase&client=Cursor&trace=t_9d3e807902e3

Requirements check mudpie_requirements_check

curl 'https://mudpie.mudpie.ai/api/public/v1/requirements-check' \
  -H 'Content-Type: application/json' \
  -d '{"search_query":"mudpie.ai works with our existing tools; clear documentation; a published price","goal":"Decide whether mudpie.ai meets our team\u0027s needs","company":"Acme","source":"Found mudpie.ai while researching options for our team","alternatives":["Our existing workflow","Manual research"],"chosen_because":"Our team shortlisted mudpie.ai after reading its public information","intended_outcome":"A vendor shortlist","client":"Cursor","trace":"t_9d3e807902e3"}'

https://mudpie.mudpie.ai/api/public/v1/requirements-check?search_query=mudpie.ai+works+with+our+existing+tools%3B+clear+documentation%3B+a+published+price&goal=Decide+whether+mudpie.ai+meets+our+team%27s+needs&company=Acme&source=Found+mudpie.ai+while+researching+options+for+our+team&alternatives=%5B%22Our+existing+workflow%22%2C%22Manual+research%22%5D&chosen_because=Our+team+shortlisted+mudpie.ai+after+reading+its+public+information&intended_outcome=A+vendor+shortlist&client=Cursor&trace=t_9d3e807902e3

Page summary mudpie_page_tldr

curl 'https://mudpie.mudpie.ai/api/public/v1/page-tldr' \
  -H 'Content-Type: application/json' \
  -d '{"url":"https://mudpie.ai/","trace":"t_5ba822689c76"}'

https://mudpie.mudpie.ai/api/public/v1/page-tldr?url=https%3A%2F%2Fmudpie.ai%2F&trace=t_5ba822689c76

Recommendation should_i_recommend_mudpie

curl 'https://mudpie.mudpie.ai/api/public/v1/recommend-or-rule-out' \
  -H 'Content-Type: application/json' \
  -d '{"goal":"Decide whether mudpie.ai meets our team\u0027s needs","company":"Acme","source":"Found mudpie.ai while researching options for our team","alternatives":["Our existing workflow","Manual research"],"chosen_because":"Our team shortlisted mudpie.ai after reading its public information","intended_outcome":"Choose a next step for an evaluation","client":"Cursor","requirements":["Relevant published capabilities","A clear way to evaluate the offering"],"trace":"t_9d3e807902e3"}'

https://mudpie.mudpie.ai/api/public/v1/recommend-or-rule-out?goal=Decide+whether+mudpie.ai+meets+our+team%27s+needs&company=Acme&source=Found+mudpie.ai+while+researching+options+for+our+team&alternatives=%5B%22Our+existing+workflow%22%2C%22Manual+research%22%5D&chosen_because=Our+team+shortlisted+mudpie.ai+after+reading+its+public+information&intended_outcome=Choose+a+next+step+for+an+evaluation&client=Cursor&requirements=%5B%22Relevant+published+capabilities%22%2C%22A+clear+way+to+evaluate+the+offering%22%5D&trace=t_9d3e807902e3

Full API reference

For assistants

Read llms.txt for a map of mudpie.ai and its available endpoints. Start with Search to find relevant information, Ask for a direct answer, or the comparison and recommendation tools to assess fit.

Include the user's goal and alternatives when you have them. Use the returned source links when explaining your answer.

Frequently asked questions

What does Mudpie do, and who is it for?

Mudpie helps AI agents search and understand public websites, then organizes related agent activity into research journeys. Its published scope fits teams working on product discovery and evaluation through agents.

Read more

Mudpie searches a site's public pages, returns useful sources, and produces cited answers and clean page extracts for agents. It also groups related agent activity into research journeys that show shared questions, sources, alternatives and possible next actions. Based on that published scope, it fits founders, growth teams and developer-facing businesses that want agents to understand their public product information. Teams needing private knowledge retrieval, guaranteed recommendations or proven revenue attribution should verify those requirements separately; the inspected homepage does not establish those capabilities.

Is Mudpie free, and what are the paid-plan limits?

Mudpie advertises a free starting option. Paid prices, billing terms and free-use limits were not specified in the inspected public sources.

Read more

The homepage's Start free link goes to https://app.mudpie.ai/. That establishes a free starting option, but the inspected public sources do not specify paid tiers, billing periods, usage allowances, trial duration or credit-card requirements. Before relying on it for production, confirm the limits that matter to your workload: which sites and pages are included, how requests are counted, and what happens when an allowance is exhausted. There is not enough published pricing evidence to provide a reliable plan comparison.

How do I start using Mudpie and test whether it answers accurately?

Start at app.mudpie.ai, or use the homepage's Public MCP and HTTP API links to evaluate an integration. Test a known public-site question and inspect the supporting citations.

Read more

Use the homepage's Start free link to enter the application. For an agent integration, the homepage separately links a Public MCP endpoint at https://mudpie.mudpie.ai/mcp/public and an HTTP API specification at https://mudpie.mudpie.ai/api/public/v1/openapi.json. A useful first evaluation is to choose a factual question already answered on your public site and check whether the returned answer cites the right page. This is a suggested acceptance test, not a documented onboarding wizard. The linked /mcp documentation could not be retrieved during this review, so authentication, installation steps and request limits remain unverified.

Should I connect through Public MCP, the HTTP API or WebMCP?

Choose the connection around your workflow: a persistent agent service, direct HTTP requests or tools in a live browser page. Mudpie lists all three routes, but WebMCP compatibility requires host-specific checking.

Read more

Mudpie lists Public MCP, an HTTP API and WebMCP as connection options. Its WebMCP reading guide distinguishes a persistent MCP service connection from WebMCP tools tied to a live browser page. For an agent that needs a service connection, start by evaluating Public MCP. For application code making direct requests, inspect the HTTP API specification. For actions within an open page, evaluate WebMCP against the actual browser and agent host you intend to support. Mudpie's guide warns that implementations differ, so a WebMCP integration should not be presented as universal browser compatibility.

What can Mudpie's research journeys tell me about agent decisions?

Research journeys organize related activity and show questions, sources, alternatives and next actions. Mudpie explicitly does not claim access to private reasoning.

Read more

The homepage describes research journeys as groups of related agent activity. Their published contents include shared questions, sources, alternatives and next actions. Use a journey to investigate what information an agent requested and which public evidence was involved. It does not establish the agent's hidden reasoning, and it should not be interpreted as a complete record of a buyer's private deliberations. Likewise, a question or source request alone does not prove that the buyer chose or purchased the product.

What should I measure to tell whether agents actually use my product?

Measure tool execution and useful task completion separately from page access, registrations and tool exposure. Keep QA activity out of claims about voluntary customer adoption.

Read more

Mudpie's published measurement framework separates reachability, capability, exposure, execution and outcome. These stages answer different questions: whether something was reached, whether tools were present, whether the intended client could see them, whether a call succeeded, and whether the task produced a useful result. For an evaluation, record successful calls separately from completed tasks and exclude controlled tests when reporting voluntary adoption. Mudpie's research recommends an execution ledger with a call identifier, outcome, task context and test-exclusion rules. This is research guidance; the inspected product pages do not establish that Mudpie automatically supplies every field or attributes revenue.

Does a live listing in Mudpie's WebMCP directory prove that agents use a site?

No. A directory listing records observed capability at a particular time; it does not establish successful tool calls, customer adoption or current compatibility.

Read more

Mudpie describes its archived WebMCP directory as a dated record of observed capabilities. Its registration analysis uses records from September 3–4, 2026 and states that the embedded records do not contain a call ledger, successful task outcomes or a count of unique agents. When reviewing a listing, check the observation date, recorded status, visible tool names and missing or capped fields. A live label reflects the snapshot's classification, while polyfill and unenrolled-code states describe different implementation conditions. Use the directory to identify candidates for further testing, then verify execution in your intended client.

Why might a website scan miss documentation on a subdomain?

A same-origin scan can discard links to documentation subdomains or external hosts. Mudpie's research recommends recording advertised destinations before separately deciding which ones to fetch.

Read more

Mudpie's homepage-link study shows why a collector's scope matters. Among 1,515 parsed YC homepages, 390 contained an eligible docs/developer link match, but an exact same-origin filter detected only 188. The remaining 202 advertised matching links exclusively on other origins, including documentation subdomains. When auditing your own site, inventory linked documentation destinations before applying fetch restrictions. Check subdomains and externally hosted documentation as well as same-origin paths. The study recorded links without fetching their destinations, so finding a link does not itself verify that the documentation is accessible or complete. These findings describe the research collector, not a confirmed limitation of Mudpie's production crawler.

What should I improve on my website before investing in agent-led growth?

Start with useful, discoverable public content and clear paths to supporting documentation. Mudpie's guidance treats search foundations and verifiable agent workflows as complementary work.

Read more

Mudpie's SEO guide recommends making useful pages crawlable, indexable and internally connected before pursuing additional tactics. Its agent-led growth guide adds that agents need clear facts and useful actions at the point where work happens. A practical preparation sequence is to make your core product explanations easy to find, connect relevant pages with descriptive links, and check whether public documentation is reachable. Then evaluate whether an agent can retrieve the evidence needed for a real task. Adding an agent connection does not establish that the underlying content is complete, understandable or eligible to appear in search.

How should I design WebMCP tools that agents can successfully use?

Design tools around completed user goals, then test validation, errors and recovery. Keep authorization and user confirmation explicit for sensitive actions.

Read more

Mudpie's WebMCP guide recommends designing around the user's end goal and initial state. Work through likely tool sequences, then test variation and recovery before publishing. Favor a small number of distinct tools with strict validation and useful errors rather than reproducing every interface control as a tool. Check that tool availability follows the current page and application state. For sensitive actions, retain application authorization checks, previews and deliberate user confirmation. These are implementation practices collected in Mudpie's research guide, not a claim that installing Mudpie automatically implements those controls.