Rejourney: Open-Source Revenue Leak Prediction for Web and Mobile Apps

Rejourney is an open-source framework designed to predict revenue leaks and user churn by analyzing real user session recordings. By combining session replay data with Large Language Models (LLMs), the tool identifies patterns of user frustration—such as rage taps and UX confusion—and suggests specific code fixes to prevent users from abandoning critical conversion events.

Automated UX Issue Detection via Session Analysis

Rejourney identifies potential revenue leaks by monitoring "critical conversion events"—such as completed signups or subscription purchases—that developers define within the SDK. The system analyzes the sequence of user journeys, including every touch, scroll, pan interaction, and rage tap, to find cohorts of users experiencing similar friction.

When a worrying trend is detected, the system processes the user recordings through an LLM (Gemini by default for cost and speed, though it supports other models like GPT-4) to analyze the touch sequence frame-by-frame. If the LLM determines that a cohort is likely to negatively impact a critical conversion event, it generates a detailed report in Markdown format. This report includes the context of the issue and a suggested fix that can be integrated into coding agents.

Technical Architecture and Data Integration

Rejourney is built to operate at scale, having been tested with approximately 2.5 million user recordings. The technical pipeline integrates several data streams to provide full context for the LLM analysis:

• SDK Support: Available for Web JS, Swift, and React Native apps.
• Contextual Data: The system bundles API response times, status codes, Application Not Responding (ANR) errors, and crash traces into the session context.
• Similarity Cohorts: A heuristic bundles user recordings into similarity cohorts to identify trends before they are sent for LLM processing.
• Code Integration: Users can optionally attach their GitHub repository, allowing the tool to include specific code fixes within the generated Markdown reports.

Privacy and Compliance

To maintain GDPR compliance and protect user privacy, Rejourney implements a strict data retention and anonymization policy. User recordings are typically retained for seven days. After this period, recordings are quantized, fingerprints are anonymized, and data is aggregated into a general analytics dashboard similar to Firebase.

Industry Perspective and User Feedback

Early adopters have reported significant improvements in conversion rates. One user reported a 30% increase in onboarding completion after two weeks of implementing fixes identified by Rejourney.

Community feedback highlights the value of an "AI-native" approach to session recording. As one user noted:

I've used LogRocket a lot... I'd love to try this as an AI native alternative which gets straight to the problem solving. It could save hours of manually watching playback and a lot of effort coming up with potential problems/solutions etc.

This indicates a shift from traditional session replay tools, which require manual review, toward automated diagnostic tools that provide actionable solutions.