You can connect iLook Face Analysis to Claude using the Model Context Protocol (MCP), and once it’s connected, Claude can run a face-shape, symmetry, or Golden Ratio read on a photo you give it, right inside the conversation — without you uploading the photo to a separate app first. This guide covers what the iLook MCP server actually does, how to connect it directly or through a gateway, and why a tool built around a photo of someone’s face deserves a different kind of care than a tool that reads text.

What the iLook MCP server actually does

iLook publishes a remote MCP server at ilook.fit/mcp that analyzes an adult photo for face shape, facial symmetry, and Golden Ratio proportions. The listing is explicit that this is consent-based analysis, not identity recognition — it’s scoring the geometry of a face in one photo, not matching that face against anyone or anything else. Prompts like these turn into tool calls once it’s connected:

  • “What face shape does this photo show, and what does that typically mean for frame or haircut advice?”
  • “How symmetrical is this face, roughly?”
  • “How close are these proportions to the Golden Ratio?”

That makes it a single-photo analysis tool, not a lookup or a database — it doesn’t store a profile, match against other photos, or tell you who someone is. See the iLook server page in gate’s directory for the current tool list and connectivity status.

Why it skips OAuth

Most servers in gate’s catalog put a consent screen in front of you because they act on an account of yours — your Airtable bases, your Qonto balance, your Dropbox files. iLook has no account to authorize against: there’s no workspace, no saved history, nothing tied to an identity on iLook’s side. Like Firecrawl and Jithox E-Invoice, it currently accepts a connection with no OAuth step at all. That can change, and a free service like this may still rate-limit heavy use, but as a starting connection there’s nothing to sign into.

The part worth sitting with is the same one that applies to every no-sign-in server: “no sign-in” describes iLook’s own authentication, not what you hand it. Here, what you hand it is a photo of an actual person’s face — a different category of data than the text or account records most tools in this series touch.

Option 1: Connect iLook to Claude directly

Claude supports remote MCP servers natively. The short version of the setup:

  • Open Claude’s connector or MCP settings.
  • Add a remote server pointing at iLook’s MCP endpoint at ilook.fit/mcp.
  • That’s it — no account to sign into before the tool shows up.

That’s enough if iLook is the only extra server you want in Claude. The tradeoff shows up once you add a second one, and then a third, each living in its own entry in your client’s config with no shared view of what’s connected or what it’s allowed to do.

Option 2: Connect it once, through a gateway

If you’re already juggling more than one MCP server or more than one AI client, wiring each one in directly gets old fast — the connection sprawl covered in what an MCP gateway is. The idea is to connect iLook (and everything else you use) to one gateway URL, then point every AI client at that single URL instead of reconfiguring each one separately.

With gate specifically, connecting iLook looks like this: add gate’s gateway URL to Claude once, then pick iLook Face Analysis from the iLook server page in the directory. From then on, the tool is available in every client connected to your gate URL, with every call going into a plain-language log instead of disappearing into a client’s own history. See the full list on the MCP servers page.

What to check before you point it at a real photo

A no-sign-in server with no account access is lower-stakes than one holding your business data, but a tool built around a face is not the same as a tool built around a spreadsheet:

  • Confirm the domain before you connect. Nothing in MCP stops someone from standing up a server that claims to be a known face-analysis tool but isn’t — and a fake one is a particularly bad place to send a real photo. Check the actual endpoint you’re pointing at — see how to tell a real MCP server from a fake one.
  • Only send a photo of someone who can consent. The listing frames this as consent-based analysis of adult photos. That instruction lives in the tool’s own description and in how you use it, not in anything MCP enforces mechanically — the protocol has no built-in way to check whose photo is in the prompt before the call goes out.
  • Read the tool description once. Even a single-purpose, no-auth server has a description the model reads before deciding what to call — in principle, a place hidden instructions could live on any server. gate’s free MCP security scanner checks a server’s tool list for prompt injection patterns before you rely on it, no signup required.
  • A geometry score isn’t a verdict. Face shape, symmetry, and Golden Ratio proportions are measurements of one photo under one set of lighting and angle conditions, not a settled fact about a person. Treat the output as a rough read, the same caution covered in how MCP tools report results — a confident-sounding number is still just a number from one tool call.
Where gate fits: gate’s directory lists each server’s currently verified connectivity — open, key-based, allow-listed, or no-sign-in like this one — so you know what to expect before you start. Every server is scanned for prompt injection before it goes live, and every call goes into a plain-language log. See the full approach at /mcp-gateway.

The bottom line

Connecting iLook to Claude with MCP takes one step, direct or through a gateway, because there’s no account behind it to authorize against. What’s worth remembering is that “no account” and “low stakes” aren’t the same thing here — the input is a photo of a real face, so the only consent check that matters is the one you do yourself before you send it, the same instinct behind treating personal data in a tool call with care everywhere else in an MCP setup.