Skip to content

The product

Find it, fix it, publish it, then check who cites it

One product with five jobs, working from the search data you already own rather than from an estimate of it.

01

The engine

What to fix, in the order it pays

Your Search Console history is mirrored every night and every URL is checked with Google's own inspection API. Rules over that data produce a ranked list, the same way every time. The model explains it; it never decides the order.

  • Pages Google has never seen, and why
  • Pages sitting just off page one, with the query they already earn
  • Pages competing with each other for the same query
  • Each finding priced by the clicks it can win, and by revenue when Analytics is connected
02

The assistant

Ask it about your own numbers

Plain language in, your own data out. It shows the steps it took and the sources it read, so an answer can be checked rather than believed.

  • Answers grounded in your data, with the working shown
  • Playbooks for the investigations people actually run
  • Follow-ups drawn from your own ranked findings
  • Any answer shareable by link, and revocable
03

Writing

Draft, rework, publish

Write a page from a finding, or rework one that already ranks without losing the queries it holds. Edit it here, publish it where your site lives.

  • A full editor, with versions and a reading measure
  • A rework is blocked if it drops a query that brings clicks
  • WordPress and Shopify directly, anything else through a signed webhook
  • Meta title and description carried through to the page
04

AI answers

Whether assistants cite you, measured monthly

Twenty buyer questions in your own language, asked of assistants that search the web, to see whether your pages are among the sources they use.

  • Claude, ChatGPT, Perplexity and Gemini, whichever are configured
  • Citation rate and mention rate kept apart, never blended into a score
  • The hosts retrieved instead of you
  • Prices quoted about you that your own pages do not support
05

Connections

Reachable from your own tools

Everything the product knows is available over an API, or from an AI client that speaks MCP, or in your inbox once a week.

  • REST API with OpenAPI docs
  • MCP server, so Claude or Cursor can read your findings
  • API keys you can revoke
  • A weekly report with the numbers that moved

Pricing

Metered in credits, not in seats you do not use

One message to the assistant is one credit, however many steps it takes. Writing an article costs more, and says so before it starts. The nightly work costs nothing.

Plans, credits, sites and seats
PlanFree$0Starter$39/moMost chosenPro$99/moAgency$299/mo
Credits a month202501,0004,000
Sites131050
Seats12520

Every plan has all of it: the engine, the assistant, writing and publishing, the AI-answer check, the API and the MCP server. The plan decides how much you can run, not what you can reach.

Annual billing costs less per month. Cancelling stops the next renewal and the current period runs to its end, as the terms say.

No account

9 tools that answer one question each

The same checks the engine runs every night, one page at a time, free and without signing in.

All free tools