Are you in the sources?See whether ChatGPT, Claude and Gemini cite your site.

Ask the questions your customers ask, on three engines with web search on, and see which answers name you and which answers cite your site. You bring your own OpenAI, Anthropic and Gemini keys, so the queries bill to your own provider accounts and there is no subscription on our side. One payment of $69 buys a founder key for life.

Get a founder key See a sample dashboard

What it tracks

Every prompt runs on each engine with web search on, under one short system prompt that asks for current information and sources. Mentions are read from the answer text. Citations are read from the engine's own citation data, never guessed from the prose.

ChatGPT
OpenAI
gpt-5-mini
Responses API with the web_search tool. A citation is a url_citation annotation on the answer text.
Claude
Anthropic
claude-haiku-4-5
Messages API with the web search tool. A citation is a web_search_result_location on the answer text. The search results the model saw are kept as well.
Gemini
Google
gemini-2.5-flash
generateContent with Google Search grounding. A citation is a grounding chunk, with its redirect resolved to the real page.

Source-set inclusion first

Being named in an answer and being in the answer's sources are two different numbers. An engine can mention your brand from memory without opening your site, and it can cite a page of yours without naming you at all. Mention rate counts the first. Source-set inclusion rate counts the second, the share of answers whose cited sources include your domain.

The dashboard leads with inclusion. The source set is where the engine actually went for its facts, and it is the number that moves when you change a page. The method comes from the experiment ledger EVOIX keeps on its own site.

A sample dashboard

One project, the latest run. Inclusion and mention rates per engine, the trend across runs, and the result for every prompt. The raw answer and the full source list for each one sit on the run page.

Sample dashboard for one project. Source-set inclusion and mention rates for ChatGPT, Claude and Gemini, a trend across runs, and a prompt table.

Your keys, your bill

Every query runs on your own OpenAI, Anthropic and Gemini accounts and shows up on their invoices, not ours. On the default models that is about a cent a query, a little more when an engine runs several searches for one answer. Treat the figure as approximate; the provider dashboards have the real numbers.

Keys are encrypted with AES-256-GCM before they reach the database, under a server-side secret that never leaves the deployment. The database holds the ciphertext and the last four characters, nothing else.

Once saved, a key is never shown again, not in Settings and not in any response the browser can read. Replace it or remove it whenever you like.

Pricing

Founder lifetime key

First 100 keys, then $149.

$69one payment, no subscription

Unlimited projects
Up to 25 prompts per project
Three engines, ChatGPT, Claude and Gemini
Daily automatic runs
Every raw answer saved

Buy a founder key

Pay once through Stripe. The key arrives by email within a minute.

Questions

Do I need my own API keys?: Yes. Every query runs on your own OpenAI, Anthropic and Gemini accounts. You paste the keys once in Settings; they are encrypted at rest and never shown again. Add one engine or all three; a run covers the engines you have keys for.
What does a run cost me?: A run sends each prompt to each engine once, billed by the providers at their rates. On the default models that is about a cent a query, a little more when an engine runs several searches for one answer, so a project with 25 prompts on three engines costs roughly a dollar a day. Those are estimates; your provider dashboards have the real numbers.
Which models?: gpt-5-mini through the OpenAI Responses API, claude-haiku-4-5 through the Anthropic Messages API and gemini-2.5-flash through the Gemini API, each with web search on. The ids are fixed per deployment and change only when a provider retires a model.
Does it track Perplexity or Google AI Overviews?: No. Version 1 covers ChatGPT, Claude and Gemini through their APIs. Perplexity and AI Overviews are not in it, and we would rather say so than show you a made-up number.
Can I export the data?: Every run page shows the raw answer and the full source list for each prompt on each engine, so nothing is hidden behind a summary. CSV export is on the roadmap.
Who builds this?: EVOIX, an AI company in DeLand, Florida. Read about the company and its AEO work at evoix.io.

Something else? Write to stephane@evoix.io.