The agent assembled a partial but useful understanding of SOMA by combining homepage metadata, an About page, and a history page — but the site's lack of accessible pricing and activity detail pages forced heavy reliance on inferred facts and web search rather than direct product information. The agent identified core value (creativity hub for ages 4–12 with strong inclusion focus) and found one concrete price point (birthday packages at €425), but could not retrieve regular workshop pricing, age-specific breakdowns, or a searchable schedule from the website itself.

• ›Step [1] (homepage) provided mission statement, target age range (4–12), and the inclusion commitment (20% free capacity) via meta tags and metadata — the site's foundational positioning was embedded in structured markup and titles, not in fetched body content.
• ›Step [4] (About page) confirmed the inclusion mission and described the space layout; Step [23] (history page) added founder context (expatriate couple inspired by California). Together these steps surfaced positioning and backstory, but not operational details like pricing or schedules.
• ›The agent attempted 10 direct fetches to logical pricing/activity pages (/pricing, /classes, /offer, /activities, /explore, /activities/for-birthdays) — all returned 404s or empty truncated HTML. The site's dynamic Next.js rendering prevented extraction of full content; the agent had to pivot to web search and external sources (venues.be, Visit Brussels) to find the €425 birthday price.
• ›Pricing for regular workshops is not published on the site itself — the agent could only infer it requires direct contact (hello@soma.brussels). The site lacks a clear pricing page or activity catalog that a machine (or human) can easily browse.
• ›The agent cited sources from the final response, but only [1], [4], and [23] were actual site fetches; the rest were web searches or external third-party listings. This indicates the site does not self-host the information needed to answer the task fully.