The agent gathered a coherent overview of Preview.io's purpose, target audience, and competitive positioning from the homepage, docs site, and a third-party review article. However, the site does not publish pricing details in any publicly accessible form — it gates this information behind a login wall — leaving a critical gap in the task requirement to explain 'how it's priced.' The agent had to work around this by inferring a credit-based model from fragments and external sources rather than retrieving it directly from the site.

• ›The homepage [1] and docs site [5] provided product positioning ('production platform for AI video,' unified workflow, collaboration features, asset management, multi-model integration, privacy-first claims) but neither exposed pricing tiers, credit costs, or trial details.
• ›The third-party review [17] (pixelsham.com) added descriptive detail about the node-based workspace and workflow ('scenes, storyboards, shot lists stay in sync') but did not contain pricing information either, indicating this is not documented anywhere in accessible web pages.
• ›The agent correctly identified that Preview.io deliberately obscures pricing behind authentication — a significant agent-readiness failure for a discovery-evaluation task. The site provides no machine-readable alternative (no pricing page, no structured data, no public tier information) and no clear freemium or trial offer to lower the barrier.
• ›Web searches [7, 9, 11, 15, 19, 21, 23, 25, 27, 29] and prior knowledge filled in competitive context (LTX Studio, Katalist, Runway, etc.) but did not yield Preview.io's own pricing, suggesting the platform has minimal public pricing documentation even in third-party reviews.
• ›The agent successfully synthesized what IS available (features, positioning, audience, differentiation) into a coherent explanation, but explicitly flagged the unsolvable pricing gap as a friction point — honest reporting of a site navigation failure rather than false confidence.