Can an AI agent find your product?

Surfex audits the machine-readable channels autonomous agents use to find, evaluate, and integrate your product. The scan is free and returns your Agent Readiness Grade. The full report is a one-time $79 purchase.

Agents do not browse your website

AI agents powering Claude Code, Cursor, OpenAI Codex, and enterprise automation platforms select products by querying machine-readable channels - APIs, structured data, agent registries, site instructions, documentation - not by browsing websites.

They do not read your blog posts, watch your demo videos, or browse your pricing page. If your infrastructure is not readable and accessible through the channels agents actually use, they will pass over you and select a competitor that is.

Most companies have never audited whether agents can find them through any of these channels. That is the gap Surfex closes.

OpenAPI Spec Not Present
MCP Registry Present but Deficient: no descriptions
CLI Scriptability No --json output
Programmatic Onboarding Browser required
Repository Readiness No AGENTS.md
A2A Agent Card Not Present
Agent Readiness Grade
D
Largely invisible to autonomous agents

Audit your product, then fix the gaps

Run the scan, read the findings, and work through the prioritized fixes in the report.

01

Audit

Enter your URL. Surfex sends real HTTP requests to your public infrastructure across six agent channels, then evaluates what comes back against the four stages of an autonomous agent's integration flow. You get a gap report with an Agent Readiness Grade and specific findings per channel. The scan reads only your public surfaces, throttled to one request per host every 1.5 seconds, and takes a few minutes.

02

Fix

The full report ranks every gap by impact and pairs it with a specific recommendation, from the OpenAPI fields agents need to the headless auth path your CLI should expose. Your team or your coding agent can work through the list directly.

Channels agents actually use

Different products expose different surfaces to agents. Developer tools live on API specs, CLIs, and repositories. SaaS products live on documentation, structured data, and llms.txt. Surfex audits each product type against the channels that matter for that category. The six dev tool channels are below, and the SaaS channel set ships next.

OpenAPI / Swagger Spec

Surfex requests your spec at twelve standard paths across your main domain and common subdomains, then hunts your docs pages and public GitHub for a linked one. If it finds a spec, it measures completeness: endpoint count, how many operations carry descriptions, whether authentication is defined and referenced, and whether examples and error schemas are present.

Programmatic Onboarding

Surfex reads your signup page, quickstart, and API-key docs and evaluates the documented path from zero to first key. It checks whether a key can be obtained without a dashboard, whether signup offers a programmatic route or requires OAuth and a browser, and whether your quickstart is executable by an agent or written for a human with a mouse.

MCP Registry

Surfex searches the official MCP registry, npm, and the public directories agents use, checks whether an mcp subdomain answers, and reads your server's published documentation to catalog its tools and how well each is described. Agent frameworks pick tools from these listings, and an unlisted or undescribed server loses to a listed one.

CLI Scriptability

Surfex searches npm, PyPI, Homebrew, and your GitHub for a first-party CLI with verified ownership, then reads its documentation for what agents need: JSON output, headless key-based authentication, and documented exit behavior. Coding agents install and drive CLIs constantly, and almost nobody audits this channel.

Repository Agent-Readiness

Surfex finds your GitHub organization, takes your most-starred verified repository, and checks it for what a coding agent looks for before contributing: AGENTS.md with build and test commands, CI configuration, dev-environment files, and contribution templates. File presence is checked against the tree, and instructions are read for substance and freshness.

A2A Agent Cards

Surfex requests the agent card at the well-known address across your domain and subdomains, validates the JSON against the required fields, and then sends one live test call to the endpoint your card declares. This is the youngest channel and the one where early publishers stand out most.

Agent Readiness Grade

Surfex derives your grade from how far an autonomous agent gets on its best available path, so the letter tells you exactly where you stand and what to fix first.

The grade is a single letter with the reasoning behind it, not a vanity score or an ambiguous percentage.

A
Agent-Ready
An agent can complete the full flow autonomously: find you, evaluate you, and obtain what it needs to integrate.
B
Substantially Ready
Agents find and evaluate you, but onboarding blocks the final step.
C
Significant Gaps
Agents can find you through some channels but cannot evaluate what you offer.
D
Largely Invisible
Agents effectively cannot find you.

Agent flow evaluation

Surfex goes further than a checklist by testing the evidence it collects against the path an autonomous agent would actually take, stage by stage, from finding you through completing a first integration. Where that evidence shows an agent would stall, the report names the stage and the signal that is missing.

A stall at a given stage is a concrete gap with a specific fix, traced back to a signal you can check for yourself.

1
Registry Presence: Is your MCP server listed in the public registries an agent would search?
2
Spec Availability: Is your OpenAPI spec published where an agent would look, and does it document authentication?
3
CLI Readiness: Does your published CLI package document headless authentication and machine-readable output?
4
Repository Readiness: Does your public repository carry the agent instructions a coding agent would need?
5
Onboarding Path: Can credentials be obtained programmatically, or does signup require a browser?

What a chat session cannot give you

Surfex runs the same instrument against every company it scans, so your grade means something relative to everyone else's. The rubric is public and the engine is versioned, which means the same domain produces the same grade twice, and a fix you ship is a change you can measure. Every finding points at a signal retrieved from your live infrastructure during the scan, and names the stage of the agent flow where it stalls.

One audit, one price

Run the free scan first. Pay once for the full report.

Free Scan
$0
See how agent-ready your product is today.
  • Agent Readiness Grade from A to D
  • Coverage Map across all six channels
  • One key finding per channel
  • Per-stage verdicts across the agent flow
Run a Free Scan

Run your free scan

The scan returns your Agent Readiness Grade and Coverage Map in a few minutes, and the full report is a one-time purchase.

Run a Free Scan