Context
The Auto-fill with AI feature uses Claude to extract metadata (amendments, participants, date, location) from a YouTube URL. However, without access to video transcripts and without the ability to have the LLM perform web searches, the extraction quality from title and description alone is too low to be useful.
Requirement
Hide the "Auto-fill with AI" button on the video submission form. The feature should not be visible to users until we have a better approach for getting video data into the system.
Notes
- The extraction endpoint (
/api/v1/videos/extract) and backend implementation can remain in place — this is a UI-only change
- The feature may be re-enabled once we have a pipeline that includes transcript analysis and/or tool-augmented extraction
Context
The Auto-fill with AI feature uses Claude to extract metadata (amendments, participants, date, location) from a YouTube URL. However, without access to video transcripts and without the ability to have the LLM perform web searches, the extraction quality from title and description alone is too low to be useful.
Requirement
Hide the "Auto-fill with AI" button on the video submission form. The feature should not be visible to users until we have a better approach for getting video data into the system.
Notes
/api/v1/videos/extract) and backend implementation can remain in place — this is a UI-only change