Whaily

User guide

How to use Whaily

Eleven steps, in the order you will do them. Every screen below is the real product.

Start free · What each feature does · MCP setup

  1. 01

    Sign in

    Use Google or an email link. There is no password to choose, and the first brand is set up from your domain.

    whaily.com/signup

    The Whaily sign-up screen: continue with Google, or send a sign-in link to an email address.
  2. 02

    Add the questions your buyers ask

    Type prompts in one at a time, or import a CSV. Each one runs on a schedule and keeps its answers.

    Prompts

    The Prompts section on its Questions tab: every tracked question with its status, the models running it, cadence, last run, and its visibility, sentiment, position and answer count.
  3. 03

    Pick the models

    Tick the models each prompt should run on. A model your plan does not include says so, and links to where you add your own provider key.

    Prompts, New prompt

    The new prompt dialog: prompt text, country, language, tags, and a model list with a reason beside every model that is not available.
  4. 04

    Read your visibility

    The share of answers that name your brand, day by day, one line per model. A day with no run is a gap, never a zero.

    Overview

    Brand visibility over time: one line per model over thirty days, with markers where a model version changed.
  5. 05

    See which model is the problem

    The same number per model and per competitor. One model can be far behind the rest, and that is the one to work on.

    Competitors, Standing

    Visibility by model: one row per model with its provider mark and version, and a column per competitor.
  6. 06

    Find the pages AI cites

    Every page a model linked to as a source, with whether it names you and which competitors it names.

    Sources

    The Sources section on its Cited pages tab: every cited page with its domain, type, whether the brand was named, competitors named alongside it, sightings, the models that cited it and its influence score.
  7. 07

    Compare against competitors

    The criteria buyers weigh in your category, scored per brand by each model.

    Competitors, Purchase criteria

    The Competitors section on its Purchase criteria tab: the criteria buyers weigh as rows, tracked brands as columns, a score in each cell.
  8. 08

    Read the answer itself

    Every answer is kept. The card shows the model, whether you were named and the brands it did name; open it for the whole answer, the order it named them in and the pages it cited.

    Prompts, Answers

    One stored answer: the model and its version, when it ran, whether the brand was named in it, and the opening prose with every brand it names in bold.
  9. 09

    Work the action plan

    The page opens on the five highest priority findings from the newest run, each as a card saying what it found and why it matters. Open one to read it in full, then accept it and it becomes an action item on the board with an owner and a due date.

    Action plan

    The Action plan section on its Recommended tab, headed Start here: five cards, each tagged High impact and saying whether it came from a check on your site or a tracked prompt, with a sentence of what it found and a link reading Open.
  10. 10

    Get told when it moves

    Set alert rules per event, to email, Slack or a webhook, plus a weekly or monthly digest.

    Settings (the gear beside your brand), Notifications

    The notifications tab: digest cadence, then alert rules by event kind with their channel and target.
  11. 11

    Connect it to your own tools

    Create a key to reach the REST API, or point Claude or Cursor at the MCP server and ask it the same questions.

    Settings (the gear beside your brand), API and MCP

    The API and MCP tab: each key with its scopes, who created it and when it was last used.

Two things worth knowing

  • A number with too few answers behind it reads N/A rather than a percentage. Whaily never rounds a small sample into a claim.
  • Being cited and being named are different. A model can build an answer on your page and name three competitors in the text; the product says which of the two it is showing.