Changelog

New updates and improvements to AnswerPeek.

New

Google AI Overviews is now a tracked engine — The answer box above Google's results joins ChatGPT, Perplexity and Gemini in every run, scored on the same prompts. Existing projects picked it up automatically — no setting to change.
Every engine now shows what it was asked, not just how it answered — Google shows an AI Overview on some queries and not others, so a prompt with no answer box is reported as exactly that — not as a prompt you lost. Each engine's card says how many of your prompts it actually answered, and prompts with no answer box are left out of the rates instead of counted against you.

New

Gemini is now tracked in every run — Google's assistant joins ChatGPT and Perplexity on every plan, queried live with Google Search grounding and scored on the same prompts. Existing projects picked it up automatically — no setting to change.

Improved

Your trend marks where the engine list changed — A third engine changes what a score is a percentage of, so the run where Gemini arrived is marked on the chart and the line breaks there. Scores either side of that mark measure different things, and the dashboard now says so instead of drawing a slope you didn't earn.

New

MCP server — AnswerPeek now speaks MCP. Connect Claude, Cursor, or any MCP client and your agent works on your live data — visibility, citations, and the fix board. Connections are listed in Settings and can be revoked there.

Improved

Sharper scoring — ChatGPT now searches the web on every prompt instead of sometimes answering from memory, and each prompt is asked several times per run. The score comes with an error bar, and a movement inside it is reported as noise rather than as news.

All of it exists to answer one question.

When a buyer asks an AI which tool to use, does it say your name? Find out in a minute, free, no signup.