Reshaping the data for a specific screen state takes seconds.
The instrumentation itself costs roughly $10 in LLM cost and 10-20 minutes of human navigation to finalize for a real-world repo.
Once instrumented: - Jump to any screen without setting up test data - Share a URL of the fully working frontend with no backend running - Collaborate with team through comments on specific items - Future: drive coding agents to make suggested changes
Do you see any other deeper business cases for this technology ?