Use case

Read a competitor's app reviews with your AI agent

Ask your agent what a rival app's users praise and complain about on Google Play or the App Store. A real run on 100 Todoist reviews.

Jesse Sibley
By , FounderPublished · Last updated · 4 min read

At a glance

JobWhat a rival app's users complain about, in their words
Toolapps_get_reviews
Credits1 for 100 reviews
StoresGoogle Play and the App Store
Run onTodoist on Google Play, 27 October 2026

The prompt

We asked our own agent, Claude Code connected to Manifold, on 27 October 2026:

TEXT
Read the 100 newest Google Play reviews of Todoist and tell me what its users complain about,
what they praise, and what a rival to-do app should do differently.

The call

StepToolInputCredits
1apps_get_reviewsTodoist on Google Play, newest 1001

The newest reviews of one app, each with its stars, text, app version, helpful votes and the developer's reply. It costs 1 credit for each 100 reviews, on Google Play or the App Store.

What came back

The 100 reviews ran from 11 September to 25 October 2026. Todoist's listing shows 4.6 stars from 306,861 ratings.

StarsReviews
560
414
36
29
111

What the 26 reviews of 3 stars or fewer were about:

ComplaintReviews
A feature is behind Pro, or the app reads as paid5
A recent update added taps to adding a task or a time4
Reminders were missing or unclear4
Hard to delete the account or start over3
Recurring tasks and subtasks2
Other, or no reason given8

The 74 reviews of 4 or 5 stars are mostly a few words, such as "great app". The longer ones praise how simple and clean the app is, that it keeps them on track (one reviewer with ADHD calls it a life saver), its recurring tasks and voice entry, its free plan, the absence of ads, and its links to Google Calendar and AI assistants.

The review with the most helpful votes, 50, gives 3 stars and says recurring subtasks lose their order when they are completed out of order.

What the agent concluded

  • The loudest complaint is an update. Four low reviews, and one 4-star review, say a recent update made adding a task or its time take more taps. Complaints that name a recent change are the ones a rival can answer in its listing.
  • Reminders are the trust problem. One reviewer says a missed reminder cost them $2,000. For a to-do app, a reminder that fires is the feature, and that is what a rival should promise and prove.
  • Users are connecting AI assistants. Four reviews name an AI assistant: two use ChatGPT with it, one found the app through Claude, one switched to it for its built-in AI. A rival without an agent connection is missing a reason people now choose.
  • Todoist answers almost everyone. The developer replied to most reviews, every 1-star review among them, often with a fix. That is the bar for a rival's own reviews.

Run it on yours

TEXT
Read the newest {number} {Google Play or App Store} reviews of {competitor's app} and group the
complaints. Which ones name a recent update?

For the listing itself and its release notes, call apps_get_app. To find the apps in a category, apps_get_top_charts returns a store's charts.

Where to go next

Questions people ask

How many reviews can I read?

Up to 1,000 a call on Google Play and 500 on the App Store, newest or most relevant first, at 1 credit for each 100.

Do I need an App Store Connect or Play Console account?

No. The reviews are public and Manifold reads them for any app, not only your own.

Why is the rating count bigger than the reviews?

Most people rate an app without writing anything. The rating count is every star rating; the reviews are the ones with text.

Jesse Sibley

Jesse Sibley

Founder of Manifold. A growth engineer who has worked with London startups for eight years.

  • 4 min read

    Find SEO quick wins in your own Search Console data

    Ask your agent for the queries where you almost rank. A real run on our own Search Console: the prompt, the calls, the output and what it changed.

  • 4 min read

    Build a B2B lead list with your AI agent

    Ask your agent for companies by what they do, where they are and their size. A real run: the prompt, the call, the list it returned and the rows to check.