Labs · Fieldwork

Experiments & Case Notes.

Fieldwork across agentic browsers, extension marketplaces, NLWeb deployments, and the MCP ecosystem — written for operators, not analysts.

⬡ Experiments in Progress ⬡
AtlasComet

Surface Trials in Agentic Browsers

Measuring citation latency, intent-to-action conversion, and policy compliance across the two leading agentic browsers. Early data: structured surfaces outperform their unstructured peers by 4–7× on first-call success.

MCPExtensions

Agentic Extension Factory

A reproducible pipeline that converts a single MCP manifest into store-ready extensions for Atlas, Comet, and Chromium peers. Build once, distribute everywhere.

NLWebJSON-LD

Citation & Answer-Box Optimization

Schema patterns that increase the probability of being cited by ChatGPT, Perplexity, and Atlas — including the entity-graph patterns that legacy SEO never had to consider.

PolicyAttestation

Signed Action Catalogs

Field tests of attested MCP manifests with scoped rate limits, human-in-the-loop hooks, and revocable trust anchors. The boring infrastructure that makes autonomy safe.

Directory · Where Presence Travels

Frameworks, Stores, and Programs We Track

  • Frameworks & SDKs
  • MCP (Model Context Protocol)
  • NLWeb (Microsoft)
  • OpenAI Apps SDK
  • Anthropic Agent SDK
  • Extension Stores
  • Atlas Marketplace
  • Comet Extensions
  • Chrome Web Store
  • Edge Add-ons
  • Affiliate & Reseller
  • Partner programs
  • White-label kits
  • Citation-share deals
  • Embedded action revenue