The agent could not access corpusprotocol.com itself (401 authentication wall), so it pivoted to researching the open-source Corpus OS project and its public documentation. It assembled a reasonably complete explanation of what Corpus OS does, who it's for, and how it differs from alternatives—but could not find pricing information or clarify whether corpusprotocol.com is a separate commercial product. The site required heavy bridge-building: public resources existed, but the actual domain in question remained closed.

• ›Step [10] (GitHub repo) and [12] (README) were the primary content sources; they returned a comprehensive technical overview of Corpus OS as a protocol suite for AI infrastructure standardization across LLM, Vector, Graph, and Embedding domains.
• ›Step [15] (corpusos.com homepage) and [18] (docs.corpusos.com) confirmed the public project exists and is well-documented, but the agent had to infer that corpusprotocol.com might be a private commercial variant—a distinction the site never explicitly surfaced.
• ›The agent had to guess and search to understand the product at all; corpusprotocol.com itself returned only '401 private preview' with no public landing page, pitch, pricing, or alternative link to public resources. The site offered zero wayfinding for unauthenticated users.
• ›Pricing information was explicitly absent from all accessible sources; the README mentions an 'open-core model' with commercial components (Corpus Router, production adapters) but no pricing page or purchasing path was found.
• ›The agent relied on GitHub README comparisons and press release references rather than on-domain marketing or positioning; the site's own web presence is minimal for a protocol suite of this scope.