curl --request POST \
--url https://{subdomain}.mihu.ai/api/v1/agents/analyze-website \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"source_url": "https://mihu.ai",
"language": "en"
}
'{
"success": true,
"message": "Website analyzed successfully",
"data": {
"company_name": "Mihu AI",
"purpose": "AI-powered contact center platform handling calls, messaging, and CRM updates.",
"industry": [
"Artificial Intelligence"
],
"services": "AI voice agents, WhatsApp automation, appointment booking, call routing",
"audience": "Contact centers, healthcare, automotive, high-call-volume businesses",
"customer_questions": "Pricing, integrations, setup time, supported languages",
"faq": "How fast can I go live? Which CRMs are supported?",
"solutions": "24/7 call answering, lead qualification, QA scoring",
"current_possible_doing_by_human": "Call answering, appointment scheduling, data entry",
"possibles_to_automate_with_ai_agent": [
"Instant 24/7 call answering"
],
"additional_analysis": {},
"contact_information": "sales@example.com",
"language": "en",
"_citations": [
"https://mihu.ai"
]
}
}Analyze website documentation for agent creation
Researches a public website or documentation URL with a web-grounded search and returns structured business context (company, services, audience, FAQ, automation opportunities) that can be used to create an agent. The URL is not fetched directly \u2014 see the note on source_url.
When to use this endpoint:
- Preview the analysis before creating an agent with /api/v1/agents/from-website
- Research a company, its services, and automation opportunities without creating anything
- Market research: add an analysis_prompt and read the result from additional_analysis
The response contains the structured analysis plus _citations with the source URLs.
curl --request POST \
--url https://{subdomain}.mihu.ai/api/v1/agents/analyze-website \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"source_url": "https://mihu.ai",
"language": "en"
}
'{
"success": true,
"message": "Website analyzed successfully",
"data": {
"company_name": "Mihu AI",
"purpose": "AI-powered contact center platform handling calls, messaging, and CRM updates.",
"industry": [
"Artificial Intelligence"
],
"services": "AI voice agents, WhatsApp automation, appointment booking, call routing",
"audience": "Contact centers, healthcare, automotive, high-call-volume businesses",
"customer_questions": "Pricing, integrations, setup time, supported languages",
"faq": "How fast can I go live? Which CRMs are supported?",
"solutions": "24/7 call answering, lead qualification, QA scoring",
"current_possible_doing_by_human": "Call answering, appointment scheduling, data entry",
"possibles_to_automate_with_ai_agent": [
"Instant 24/7 call answering"
],
"additional_analysis": {},
"contact_information": "sales@example.com",
"language": "en",
"_citations": [
"https://mihu.ai"
]
}
}Authorizations
Use a Bearer token to access these API endpoints. Example: "Bearer {your-token}"
Body
Website or documentation URL the analysis is about. NOTE: the URL is not fetched. It steers a web search over already-indexed pages, so a deep link is a hint rather than a page that will be read, a URL that 404s still returns an analysis instead of an error, and a site with a thin search footprint returns generic content or a same-named different company. Check the _citations in the response against this host before trusting the result. Required unless analysis_prompt is provided.
"https://example.com/docs"
Optional output language (ISO 639-1, e.g. en, tr, fr, de). All analysis fields are written in this language. Defaults to the workspace language.
"fr"
Optional extra analysis instruction. Use it to request pricing analysis, documentation summary, or market positioning on top of the standard analysis. The result is returned inside additional_analysis. Can also be used alone (without source_url) for prompt-only research, but results are better when a URL anchors the analysis.
4000"Summarize the product documentation, pricing pages, and FAQ in detail."
Optional ISO 3166-1 alpha-2 country code (NL, US, DE, FR, GB...). Localizes the web search to that market. Use when the question or company is market-specific.
"NL"
Optional state, province, or administrative region for finer localization. Use together with search_country.
"North Holland"
Optional city for local-business analysis. Use together with search_country.
"Amsterdam"
Optional latitude for precise location-based search. Use with search_longitude and search_country.
41.0082
Optional longitude for precise location-based search. Use with search_latitude and search_country.
28.9784
Optional ISO 639-1 language codes. Restricts which source languages are searched — e.g. ['nl'] to analyze only Dutch sources. Independent from the output language.
10Optional domain allow/deny list, maximum 20 entries. Plain domain includes it ('mihu.ai'), '-' prefix excludes it ('-wikipedia.org'). Use to keep the analysis pinned to the company's own sites or exclude noisy sources.
20Analysis depth. When to use each value:
- fast-search (default): fastest and lowest cost. Small websites, quick summaries. Typical response in seconds.
- pro-search: balanced speed and quality with better page coverage. Recommended for most business analysis.
- deep-research: slow, deep crawl and research. Large or complex websites, or extra research via analysis_prompt. Can take minutes.
- advanced-deep-research: slowest and most exhaustive. Use only when maximum coverage matters more than speed.
fast-search, pro-search, deep-research, advanced-deep-research "pro-search"