Foliora MCP¶

Hosted remote MCP server (Streamable HTTP, no auth at v0, read-only by design) for Foliora's managed AI-search service. Foliora improves the public sources a company owns so it can be found on Google and cited by ChatGPT, Gemini, Claude, Grok and Google AI Overviews. Its MCP surface is deliberately narrow and safe: an agent can read what Foliora is, hand a human a preview link, or fetch a snapshot the human already started. It cannot crawl, take credentials or publish - humans still start the work.

Server type: Remote (Streamable HTTP)
Auth: None for v0
Endpoint: https://www.foliora.ai/mcp
Registry: ai.foliora/search (0.3.0)
Tools: 3 (all read-only)
Pricing: Free preview; see foliora.ai/pricing
Docs: https://www.foliora.ai/mcp · Skill: foliora.ai/skills/foliora/SKILL.md
Category: SEO

Why This Matters for Operators¶

AI-search visibility (GEO/AEO) is becoming a revenue channel, but the work still needs a human to kick off a crawl and review findings. Foliora's MCP makes the agent the front door: it explains the service, creates a /preview link for any public website, and reads back the snapshot once the human starts it in a browser. The guardrail is the feature - the agent can never crawl, hold credentials or publish, so handing an agent a Foliora connection is risk-free while the human does the judgment work. Operators get GEO/AEO previews into their existing agent workflows without standing up audit infrastructure.

Tools & Capabilities¶

Tool	Purpose
`get_product`	Foliora product facts: brand, plans, canonical URLs and what the server will not do
`get_snapshot`	Fetch a preview snapshot the human already started (32-character hex token from /preview/{token}); public findings and pages only
`create_preview_link`	Build a Foliora /preview URL for a public website and send the human there to start the snapshot

REST names match MCP tool names exactly; the same handlers serve /api/v1.

Installation¶

claude mcp add --transport http foliora https://www.foliora.ai/mcp

No key, no signup for the MCP surface. The human completes previews in the browser; the agent polls get_snapshot for the results.

Configuration¶

{
  "mcpServers": {
    "foliora": {
      "url": "https://www.foliora.ai/mcp"
    }
  }
}

Business Relevance¶

Founders and marketing leads get AI-citation previews into the agent they already use daily
SEO teams hand a preview link to stakeholders instead of exporting reports
Agencies let clients self-serve previews while the agency runs the actual GEO program
Operators new to GEO get a zero-risk read-only entry point - the agent cannot change anything

Integration with CorpusIQ¶

Foliora previews AI-search visibility; CorpusIQ measures what that visibility earns. An operator can run a Foliora preview on a landing page, apply the fixes, then track the resulting traffic and revenue through CorpusIQ's GA4 and Google Ads connectors - closing the loop from "are we citable" to "did citations move revenue". For CorpusIQ's own GEO program (docs.corpusiq.io), a Foliora snapshot on the docs domain gives a second read on AI-citability alongside the platform's internal checks.

Limitations¶

Brand new - v0 registry version, no public adoption track record yet
Read-only by design: the agent cannot start a crawl itself, so workflows are human-gated
Preview snapshots only - no full audit API over MCP yet
No auth at v0 means the surface is product-facts plus preview plumbing, not account data