AI Visibility MCP for Windsurf

In Windsurf, the raw answers the six engines actually give are one Cascade request away: ask, and fetch_raw_answers returns the answer text plus every cited source as a tool result. One MCP server, one deliberately narrow tool - and Cascade's agentic flow chains the rest: fetch, diff, edit, all from a single prompt.

Connect AgentGEO to Windsurf

Windsurf configures MCP servers in JSON. Add an agentgeo entry to ~/.codeium/windsurf/mcp_config.json:

{
  "mcpServers": {
    "agentgeo": {
      "command": "npx",
      "args": ["-y", "agentgeo-mcp", "--key", "ag_live_..."]
    }
  }
}

The server is a zero-dependency npm package that speaks MCP over stdio (Node.js 18+). A live ag_live_ key comes with a plan. Trying the wiring first? An ag_test_ key always returns clearly labelled demo records at zero credits. Reload Windsurf, and fetch_raw_answers becomes an ability Cascade can call in a flow.

Each call returns one record per surface - the provider's answer, verbatim, with every source it cited as structured data (the full record also carries the raw providerFields):

{
  "surfaceKey": "perplexity",
  "status": "delivered",
  "answerText": "...",
  "sources": [
    { "title": "...", "url": "https://...", "position": 1 }
  ],
  "fetchedAt": "2026-07-15T09:12:04Z",
  "latencyMs": 8412,
  "providerRecordId": "..."
}

What to ask once it's connected

No arguments to memorize - describe the outcome and Cascade chains the steps. Say "in Spain" and it passes country through (it defaults to US). Worth knowing: language only reaches Google AI Overviews, where it drives the search locale - the chatbot surfaces take a country and no language field. Asks that let one flow do the whole job:

• "Fetch what ChatGPT and Perplexity answer for 'best email API for developers' and list which domains each cites."
• "Run our homepage query across all six surfaces, diff the citation lists, and if ourdomain.com is missing from Gemini, edit the hero copy to answer the query head-on."
• "Pull Perplexity's answer for our category, compare sources[] to geo-log.md, and open the docs page that lost its citation for editing."
• "Diff google_ai_overview against google_ai_mode for 'best observability tools' and tell me what AI Mode adds."
• "Fetch Copilot's answer for 'alternatives to our product' - are we in the answer text, or only the sources?"
• "Sweep all six engines for our top query, then draft the landing-page changes that would earn the citations we're missing." One flow, fetch to fix.

What the MCP server gives you

Deliberately thin - an answer-access layer for Cascade, not an analytics product bolted into your editor.

• One narrow tool - fetch_raw_answers posts to AgentGEO's API and returns the run envelope verbatim. Nothing else to learn, nothing else to break.
• Six engines behind one contract - ChatGPT, Perplexity, Gemini, Google AI Overviews, Google AI Mode and Copilot with an identical record shape, so widening a check is one more word in the ask.
• Structured sources[] - every citation with its title, URL and position, ready for Cascade to diff, count or grep against your domain. No links regexed out of prose.
• Raw records only - no rankings, sentiment or visibility scores. The conclusions come from Cascade, with your repo and history in context.
• Managed-scraper engine, maintained for you - under the hood the server talks to POST https://api.agentgeo.org/v1/fetches; keeping six collection paths healthy is AgentGEO's job, not a headless browser in your project.
• Usage-based billing - one credit per run, failed runs cost zero, priced by usage rather than per seat.

What it doesn't do

The narrowness is the point. Cascade holds your repo, your positioning and your judgment; conclusions computed outside that context would be worth less than raw records analyzed inside it - that is the whole GEO data layer argument. So, honestly:

• No dashboards. Records arrive as tool results; the report is whatever Cascade writes from them, committed in your repo.
• No visibility scores or sentiment. Raw answerText and sources[] only - the judgment layer is deliberately left to your agent.
• No recommendations engine. Cascade drafts fixes against your actual pages, which beats generic advice every time.
• Not instant. Requests wait up to 180 seconds - live surfaces are slow. An AI Overview SERP round-trip runs 40-90s, and some chatbot scrapes exceed the sync budget entirely: those records come back failed with a providerFields.snapshot_id, and a follow-up call with that id and the same single surface redeems the finished answer without paying for a re-scrape (not valid for google_ai_overview). A flow that fetches several surfaces will pause on the slow ones.

Prefer the REST API?

The MCP server is a thin wrapper over the same endpoint your code can hit directly: POST https://api.agentgeo.org/v1/fetches. Script the sweep in Python or curl and you get the identical contract - same surfaces, same records, same credits - so a check born in a Cascade flow can graduate to CI unchanged.

Who builds on this

Developers working product sites and docs in Windsurf who want Cascade to own the whole loop. When one prompt can fetch how the six engines answer for your category, diff the citations and edit the page that's missing, the agentic flow is doing GEO work end to end instead of handing you a report to act on. See where you stand first with the free audit, or see how the approaches compare.