# Recomma docs

> Recomma measures how AI models describe a brand: what they say, which sources they lean on, and what closes the gap.

## Pages

- [Concepts](https://docs.recomma.ai/concepts): Every word the product uses — project, prompt, chat, sample size, confidence interval, source, citation — and exactly what each one counts.
- [Welcome to Recomma](https://docs.recomma.ai/): What Recomma measures in ChatGPT, Gemini and Google's AI answers, how every figure is counted from real chats, and what it takes to start.
- [Metrics overview](https://docs.recomma.ai/metrics): Four figures, what each one asks, and when a reading is worth acting on.
- [Market value](https://docs.recomma.ai/metrics/market): What the questions you track are worth, how much of that market they reach, and how much of it names you.
- [Position](https://docs.recomma.ai/metrics/position): How early in a chat your brand is named among the tracked brands, averaged over the chats that name you — the one figure that rewards being first.
- [Sentiment](https://docs.recomma.ai/metrics/sentiment): How favourably a chat speaks about your brand, scored 0–100 and averaged over the chats that name it — what the scale means and who scores it.
- [Share of voice](https://docs.recomma.ai/metrics/share-of-voice): Your share of all the brand mentions once every tracked brand is counted — how it differs from visibility, and why repeating a name does not help.
- [Visibility](https://docs.recomma.ai/metrics/visibility): The share of chats that name your brand at least once — the headline figure, what its denominator counts, and the edges it leaves out on purpose.
- [Agent search](https://docs.recomma.ai/product/agent-search): Where the search indexes an agent calls put your brand, which is not where the assistants put it.
- [Competitors](https://docs.recomma.ai/product/brands): The field you are measured in: the tracked competitors that share of voice and position are counted across, and your own row among them.
- [Chats](https://docs.recomma.ai/product/chats): The AI responses every figure in the product is measured from — what each chat records, and how to trace any number back to the text behind it.
- [Impact](https://docs.recomma.ai/product/impact): Whether the thing you did moved the number: where the brand's visibility is heading, and a verdict on each action once it has been re-measured.
- [Keywords](https://docs.recomma.ai/product/keywords): The priced market behind your prompt set, row by row, and the demand no question measures.
- [Opportunities](https://docs.recomma.ai/product/opportunities): The sites already trusted on your subject that say nothing about you.
- [Overview](https://docs.recomma.ai/product/overview): The page you land on, panel by panel: the filter bar, the visibility trend, rankings, the priced market, top cited domains and the latest chats.
- [Prompts](https://docs.recomma.ai/product/prompts): The questions you track, and everything that follows from choosing them well.
- [Ranking](https://docs.recomma.ai/product/ranking): The whole field on one page, ordered by whichever of the four figures you are asking about.
- [Site](https://docs.recomma.ai/product/site): Your own site as the engines meet it — what was crawled, what the markup says, and who is allowed to read it.
- [Sources](https://docs.recomma.ai/product/sources): Which sites the models trust on your subject, and what they say about you.
- [Traffic](https://docs.recomma.ai/product/traffic): The people an AI assistant actually sent to your site, from your own Google Analytics — who sent them, where they landed, and whether they converted.
- [Your brand](https://docs.recomma.ai/product/you): Your own brand's page: how each figure moved, where you are named and where you are not, by model, and the sources behind each.
- [Quickstart](https://docs.recomma.ai/quickstart): From a bare domain to a first reading you can trust: add the brand, fix the prompts and competitors it proposes, let it sample, then read.
- [MCP server](https://docs.recomma.ai/reference/mcp): Connect Claude Code, Codex or any MCP client to your workspace, and let it read what Recomma measures.
- [Models](https://docs.recomma.ai/reference/models): The surfaces Recomma asks, the ones it does not, and why how it asks them matters.
- [How sampling works](https://docs.recomma.ai/reference/sampling): What one sampling cycle does, what it costs in chats, and how stored chats become readings with a sample size and a confidence interval.