Skip to Content
Connecting AI assistants

Connecting AI assistants

AI Connections let you connect an AI assistant — Claude Code, Claude Desktop, or any MCP  client — directly to your Voris analytics. Once connected, you can ask questions about your traffic, conversions, and AI visibility in plain language, and the assistant reads the answers straight from your data.

This page explains what a connection does, how to create a key, and how to connect each client.

What an AI Connection does

Voris runs a hosted MCP server at https://mcp.voris.ai/mcp. MCP (the Model Context Protocol) is the open standard AI assistants use to talk to outside tools. Connecting your assistant to this server gives it a set of read-only tools for one of your sites:

  • Metrics — visits, visitors, and trends over any date range.
  • Top pages, sources, and countries — your busiest pages and where traffic comes from.
  • Conversions — goal completions, broken down by source, country, or device.
  • AI visibility and citations — how AI engines see and cite your site.
  • Tracking plan — the events your site is set up to collect.
  • Ask the analyst — a natural-language question answered from your data.

Each connection key gives one AI client read-only access to one site. A key can never change your data, your settings, or your billing — it can only read analytics. You can revoke any key at any time.

Add Voris as a custom connector in Claude (OAuth)

If you use Claude.ai (web, desktop, or mobile), the quickest way to connect is to add Voris as a custom connector — with no connection key to copy. You sign in with your Voris account and Claude connects over OAuth.

  1. In Claude, open Settings → Connectors → Add custom connector.
  2. Enter a Name (for example, “Voris”) and set the Remote MCP server URL to https://mcp.voris.ai/mcp.
  3. Open Advanced settings and set the OAuth Client ID to anthropic-claude-connector. Leave the OAuth Client Secret empty — Voris is a public client, so there is no secret.
  4. Click Add, then open the connector and click Connect. Claude sends you to Voris to sign in and choose which site to share, then returns you — connected.

Like a connection key, a custom connector is read-only and scoped to one site. To connect a different site, add another connection and pick that site there.

No Voris account yet? You can create one during the connect flow. If you don’t have a Voris account — or your account has a workspace but no site yet — connecting from Claude walks you through creating your account, workspace, and first site inline, then returns you connected. There’s nothing to set up in advance. A brand-new site has no analytics until you install tracking on it, which you can do any time after connecting.

Create a connection key

  1. Open Settings and go to AI Connections.
  2. Click Generate connection key.
  3. Give the key a name (for example, “Claude Desktop on my laptop”) and choose the site it should read.
  4. Click Generate key.
  5. Copy the key now. For your security it is shown only once — if you lose it, revoke it and generate a new one.

The dialog also shows a ready-to-paste snippet for each client. Keep your key somewhere safe, like a password manager, and never commit it to a public repository.

Connect Claude Code

Claude Code speaks remote MCP natively. From your project directory, run:

claude mcp add --transport http voris https://mcp.voris.ai/mcp --header "Authorization: Bearer <your-key>"

Or add it to your project’s .mcp.json by hand:

{
  "mcpServers": {
    "voris": {
      "type": "http",
      "url": "https://mcp.voris.ai/mcp",
      "headers": { "Authorization": "Bearer <your-key>" }
    }
  }
}

Voris is available in your next Claude Code session — try asking “How many visitors did my site get last week?”

Connect Claude Desktop

Claude Desktop cannot reach a remote MCP server directly — it can only launch local programs. To bridge the gap it uses mcp-remote, a small helper run through npx, so you need Node.js  installed.

  1. In Claude Desktop, open Settings → Developer → Edit Config. This opens claude_desktop_config.json.

  2. Add the voris server (merge it into any mcpServers you already have):

    {
  "mcpServers": {
    "voris": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://mcp.voris.ai/mcp",
        "--header",
        "Authorization:${AUTH_HEADER}"
      ],
      "env": { "AUTH_HEADER": "Bearer <your-key>" }
    }
  }
}

  3. Save the file and restart Claude Desktop completely.

The key is passed through the AUTH_HEADER environment variable rather than written straight into the arguments — that keeps the space in Bearer <your-key> from splitting the header. Once Claude Desktop restarts, you will see voris in its tools.

Connect another MCP client

Any MCP client that supports Streamable HTTP can connect. Point it at the endpoint and send your key as a bearer token:

Endpoint:      https://mcp.voris.ai/mcp
Authorization: Bearer <your-key>

What you can ask

Once connected, ask your assistant in plain language, for example:

  • “What were my top 10 pages last month?”
  • “Did conversions go up or down week over week?”
  • “Which AI engines cited my site, and for which prompts?”
  • “Summarise this week’s traffic and flag anything unusual.”

Revoking a key

Revoke a key the moment a laptop is lost, a key is shared by mistake, or a client no longer needs access:

  1. Open Settings → AI Connections.
  2. Find the key in the list and click Revoke.
  3. Confirm with Revoke key.

The key stops working immediately, and any AI client using it loses access on its next request. Revoking one key never affects your other keys.

How to add an llms.txt to your sitePrivacy modes