State of AI-Built Apps
ShipForge is building the first original dataset on how AI-built mobile apps actually perform at launch readiness. This page publishes aggregate findings — never individual app data, and never a statistic we cannot back with real scans.
In data collection
Findings below are not yet published. They appear here only after enough real Launch Score runs and Launch Doctor scans exist, and after a privacy review confirms no app or owner is identifiable.
What we plan to measure
Each metric below is computed from anonymous, consent-safe aggregates only. We will publish a number only when real data supports it; until then it stays blank.
- Apps with exposed secrets / hardcoded API keys—pending data
- Apps missing account deletion—pending data
- Apps missing crash / error monitoring—pending data
- Apps with subscription / entitlement gaps—pending data
- Average first Launch Score—pending data
- Median score improvement after rescan—pending data
- Most common launch blockers—pending data
How the data is collected
Every Launch Score is a self-assessment the developer runs in their browser. A Launch Doctor Review is an evidence-backed automated scan of a real repository. For research, we aggregate only: which checklist categories pass, warn, or fail — never repository names, file paths, secrets, source code, owner identity, or customer results. Individual scans are never published.
Why this matters
Most advice for AI-built apps is anecdotal. A real, growing dataset on where these apps fail before launch — and how much they improve after targeted fixes — is the most useful thing ShipForge can give back to the builder community. When the first report is ready, it will be free and citable.
Get notified when the first report ships
The dataset grows with every Launch Score. Run yours to contribute an anonymous data point and get the report the moment it is published.