All the Possibles

We build the checks agents run on themselves.

Small, focused tools: each one either catches a real problem before it ships, or gives an agency one dashboard instead of twenty logins.

Preflight

MCP security

Check it before your agent trusts it.

Scans an MCP server or a page/document for unauthenticated transport, capability overreach, and injected instructions aimed at the model, before your agent trusts either one.

Live: free, Pro from $19/mo Open Preflight

Design Check

Design critique

Does it read as AI-generated?

Checks a page's layout, typography, and copy against the patterns that make AI-built work look and sound generic, before you ship it, not after a reviewer flags it.

Live: free, Pro from $19/mo Open Design Check

Vantage

SEO reporting

Every client's Search Console, one dashboard.

A reporting console built for agencies running search for more than one client: no more logging into each property's Search Console separately.

Live: $39/mo per client Open Vantage

Bearing

Meta Ads reporting

Every client's ad account, one dashboard.

Sits above Meta Ads Manager for agencies running more than one client's spend: every account's results and what needs attention, on one screen instead of switching Business Managers.

In build, not yet accepting clients Open Bearing

Pulse

Core Web Vitals

Real field data, not a lab guess.

Checks a page against Google's actual ranking-signal data (Chrome UX Report, real visitors) and a live Lighthouse run, then says plainly whether it passes. Free, no signup.

Live: free checks, Pro monitoring from $19/mo Open Pulse

Verdict

Design review

A design review before the design ships.

Six review tools an agent can call on a screenshot, a flow or a spec: hierarchy, copy, states, accessibility, the AI-generated tells, and a verdict with the fix list. Bring your own Anthropic key or take Pro.

Live: two demo calls a day, Pro from $19/mo Open Verdict

Toolbox

Small tools

Fourteen mundane jobs, done properly.

AI-writing tells, the token cost of an MCP tool list, CLAUDE.md lint, install snippets, AI-crawler policy, security headers, page meta, JSON to schema, cron, an env scrubber, a redirect tracer, a sitemap check, an MCP tool-definition lint and a JWT decoder. Free, no signup.

Live: free, Pro $5/mo Open Toolbox

Designers

For designers using AI

The parts of design your AI gets wrong, done properly.

Connect one web address to Claude, Cursor, VS Code, Codex or ChatGPT and it can check contrast exactly, audit a page for accessibility, run a Nielsen heuristic evaluation, spot generic AI design and writing, plan research and critiques, and write handoff specs. Thirteen free skills to match. Works with a Figma link; nothing to install.

Live: free, Pro $5/mo Open Designers

Design Skills

Design skills

Thirteen free skills, no signup.

Design critique, accessibility review, Nielsen heuristics, an anti-slop check, UX writing, research plans, and seven more. Install the plugin in Claude Code or copy one skill folder into your own project, then point it at a screenshot, a flow, or a PRD.

Live: free, no signup View the skills on GitHub

Preflight and Design Check share the same mechanic: an agent calls a check before it ships or trusts something, and gets told what's wrong before a person ever sees it. Vantage runs on a different clock: real subscription revenue for a real buyer, an SEO agency, while the agent-tooling side builds an audience.

Membership, $15 a month

Pro on every product: Preflight, Design Check, Pulse, Verdict and Toolbox. One subscription, one invoice, five API keys emailed together. Five Pro plans for less than three. First seven days free, cancel from one page. Yearly is $150, two months free.