Context
OpenAI opened app submissions for review and publication in ChatGPT and documented an in-product app directory where users can browse or search for published apps.
What changed
Approved third-party apps gained a documented ChatGPT discovery surface. OpenAI also said it was experimenting with surfacing relevant apps directly in conversations using signals such as conversational context, app usage patterns and user preferences.
Why it matters
This establishes an agentic discoverability layer inside ChatGPT that is distinct from ordinary web retrieval and citation visibility.
Visibility OS interpretation
Visibility OS treats ChatGPT app discovery as a separate distribution surface. The disclosed experimental signals should not be interpreted as a complete or fixed ranking algorithm.
What remains uncertain
How will conversational app recommendations be ranked as the system matures? Which directory metadata or quality signals materially affect discovery? How stable are recommendation signals across users and contexts?