Automatically classify app store reviews: bug reports vs feature requests
Monitr by MRVL uses machine learning to classify every app store review as a bug report, feature request, positive feedback, or noise, then routes each signal to your team's tools in real time.
How Monitr classifies reviews automatically
Every review from the App Store and Google Play is ingested by Monitr and tagged by its ML classifier into one of five categories: bug_report, feature_request, pr_crisis, positive_feedback, or noise. This happens within minutes of publication. The classifier learns from thousands of real app reviews, so it recognises patterns that manual tagging would miss. A genuine crash report gets flagged differently from a user wishing for a dark mode toggle. Once classified, each signal moves to the next step: routing to your team.
Route bugs and features to the right place
After classification, Monitr routes signals automatically. Bug reports can go to Linear, Jira, or GitHub Issues for your engineering team. Feature requests can land in Shpd, a public roadmap tool, so your community votes on what matters most. Crisis alerts hit Slack every 15 minutes if sentiment spikes. All this happens without manual intervention. You set the rules once, and Monitr enforces them across every app you monitor. If you use multiple tools, Monitr integrates with Slack, Linear, Jira, GitHub Issues, Trackr, and Shpd. As of June 2026, Monitr processes over 5,000 mentions per month for Studio plan users.
Correlation detection groups related bugs into patterns
A single one-star review is noise. Ten reviews saying the app crashes on iOS 18.2 within an hour is a pattern. Monitr's hourly correlation detection groups related signals into narratives, so you see the real story behind the noise. If fifty users report the same bug in different words across the App Store and Reddit, Monitr surfaces that as a single correlated incident. This saves your team from chasing false alarms and helps prioritise genuine blockers.
Monitor your apps and your competitors
Monitr watches five sources: App Store reviews, Google Play reviews, Twitter and X mentions, Reddit posts, and Google News. The free plan covers one app with up to 50 mentions per month. The Studio plan (£99/mo) covers five apps. The Pro plan (£299/mo) covers 20 apps and adds competitor app monitoring, so you stay ahead of what users are saying about rival products. Portfolio plan (£599/mo) covers unlimited apps, includes a REST API, and gives you full programmatic access to classification data.
Set up classification rules without code
Monitr works through a web dashboard. You connect your app store accounts and Twitter/X bearer token, then define routing rules: any bug flagged as critical goes to Linear, feature requests go to Shpd, crisis alerts go to Slack. The ML classifier runs on every new mention, and your rules execute automatically. If you need custom logic or want to pull classification data into your own systems, Portfolio plan customers get REST API access to the full signal stream.
Weekly digests keep stakeholders informed
Beyond real-time routing, Monitr sends a weekly digest email summarising all signals: how many bugs were reported, which features users asked for most, sentiment trends, and any crisis alerts. This gives product managers, marketers, and executives a snapshot of brand health without logging in. The digest is configurable, so you control which apps and metrics appear in each recipient's email.
Ready to try Monitr by MRVL?
One tap to download. No sign-up wall.
Frequently asked questions
What counts as a bug vs a feature request in Monitr's classifier?
A bug report typically mentions crashes, errors, broken functionality, or unexpected behaviour. A feature request asks for new capability, UI changes, or settings. Monitr's ML classifier learns these patterns from thousands of real reviews and tags each signal accordingly. You can also review and retrain the classifier via the dashboard if it misclassifies a signal.
Can Monitr reply to reviews automatically?
No. Monitr is a monitoring and classification tool, not a social media management platform. It ingests reviews and classifies them, then routes signals to your team or tools. Your team decides how to respond. This design keeps the decision-making human and prevents automated replies that miss context.
How quickly does Monitr classify new reviews?
Classification happens within minutes of a review being published on the App Store or Google Play. Crisis alerts are sent every 15 minutes if sentiment shifts significantly. Correlation detection runs hourly, grouping related signals into narratives so you spot patterns fast.
Which tools can Monitr route classified signals to?
Monitr routes signals to Slack, Linear, Jira, GitHub Issues, Trackr, and Shpd. You define routing rules based on classification tags, so all bugs go to one tool, feature requests to another, and crises to your on-call channel.
Can I monitor competitor apps with Monitr?
Yes, but only on the Pro plan (£299/mo) and above. Competitor monitoring tracks the same five sources (App Store, Google Play, Twitter/X, Reddit, Google News) for rival apps, helping you stay aware of what users say about competing products.
Does Monitr monitor private messages or private Slack channels?
No. Monitr monitors only public sources: App Store reviews, Google Play reviews, public Twitter/X posts, public Reddit threads, and Google News. It does not access private messages, private Slack channels, or other gated conversations.