What is MentionFlow
MentionFlow shows you how AI assistants talk about your brand, and what to change. Here is what it does and how these docs are laid out.
Someone asks ChatGPT, Perplexity, or Gemini "what's the best tool for X?". Your brand is in the answer or it isn't. Until now you had no way to see which.
MentionFlow is an SEO platform for AI assistants instead of Google. It asks the engines the questions your buyers ask, over and over, and saves what they say back. Then it finds every place your brand and every competitor is named or cited.
So you can see how visible you are in AI answers, why you rank where you do, and what to change. Every number is backed by the real answers, so you can check the receipts.
Who these docs are for
- People who use MentionFlow: marketers, founders, and agencies who want to read a metric, set up a project, or understand a report.
- People who build on MentionFlow: developers using the v1 API or the MCP server, and engineers working on the product.
Every feature page has the same shape: what the screen does, how to use it, how it computes what it shows (denominators, windows, rounding), its limits, and the related API endpoints. The "how it computes" sections are checked against the code.
The model in one minute
A few nouns carry the whole product. Core concepts defines them all. The short version:
- A brand (also called a project) is what you track. A workspace owns one or more brands and holds your plan.
- Each brand's collection config sets how it is sampled: its markets (up to five), a set of engines, and a set of prompts, on a schedule. Edit the market, persona, and the project's engines in Collection. New projects use the workspace default engine set from Engines until you change them.
- A prompt is a question you want to watch, for example "best project management software for agencies".
- A run is the smallest unit of data: one prompt, asked to one engine, in one region, on one day. Each run stores the answer word for word.
- A receipt is a finished run shown in full: the answer text, the ranked brand mentions, the citations, and any shopping or ad surfaces.
Visibility, share of voice, sentiment and sources are all built from runs over a rolling time window.
MentionFlow samples the consumer versions of these engines (logged out, geo-targeted), not their developer APIs. That is what your customers actually see. The Methodology page explains how each engine is sampled and why.
How to read this section
- New here? Start with the Quickstart. About five minutes to a first collecting project.
- Want the vocabulary first? Read Core concepts.
- Picking a plan, or hitting a limit? See Plans and quotas.
- Building an integration? Jump to the API and MCP reference.
Prefer to learn inside the product? The in-app Learn academy walks you through each screen. These docs are the reference above it.