Anthropic Ships MCP Registry as ChatGPT Deepens Operator API Reach
Two structural shifts landed this week: Anthropic's official MCP server registry went live, giving agents a verified directory to discover and connect business tools, while OpenAI expanded its Operator API to additional enterprise tiers — meaning the pipe between AI agents and your business just got more formal, and unregistered businesses are now actively invisible to a growing class of autonomous queries.
This Week's Signals
Anthropic Launches Official MCP Server Registry, Raising the Stakes for Business Discoverability
Why it matters for your score
The MCP registry is the first vendor-maintained, crawlable index of businesses and tools that Claude agents can discover and invoke autonomously — without a human typing a search query. If your business isn't listed with a conformant MCP server manifest, you are structurally absent from this discovery layer. Businesses that publish a valid MCP endpoint this week get first-mover indexing advantage before the registry fills with competitors.
OpenAI Extends Operator API Access to Mid-Market Enterprise, Enabling Agent-Initiated Transactions
Why it matters for your score
OpenAI's Operator API now lets ChatGPT agents browse, select, and transact with businesses on behalf of users — previously limited to a handful of design partners, now open to a broader enterprise tier. Businesses that expose structured pricing, availability, and checkout flows in machine-readable formats will be preferentially selected by Operator-driven workflows. If your pricing page is only human-readable, the agent will skip you for a competitor who published a JSON pricing schema.
Perplexity Confirms Citation Weighting Now Favors Pages With Validated llms.txt and Structured FAQs
Why it matters for your score
Perplexity's engineering blog confirmed that its June citation-quality update applies a positive ranking signal to domains that serve a valid llms.txt file and contain FAQ or HowTo schema markup — treating them as higher-authority sources for answer synthesis. This is the first public confirmation from a major AI answer engine that llms.txt directly influences citation selection, not just crawler access. Any business generating leads from Perplexity recommendations needs both files live before the next crawl cycle.
What to do this week
- 1
Register your MCP endpoint this week while the Anthropic registry is newly open and competition for indexed slots is low — use Faro's /tools/ai-readiness-scan to confirm your server manifest is conformant before submitting, then cross-reference your tool descriptions against the registry's required metadata fields.
- 2
Publish or update your llms.txt file immediately to capture Perplexity's new citation-weighting signal — Faro's /tools/llms-txt generator will output a spec-compliant file in under two minutes; prioritize including your service categories, geographic coverage, and a plain-language capability summary, as these are the fields Perplexity's update specifically references.
- 3
Audit your pricing and service pages for machine-readable structure before OpenAI's Operator API routes transactional queries at scale — run Faro's /tools/ai-schema tool to add JSON-LD Offer and Service schema, and verify your robots.txt explicitly permits GPTBot, ClaudeBot, and PerplexityBot using /tools/robots-txt so agent crawlers aren't blocked at the door.