Upload + processing UI — drag & drop, format detection, progress screen

[Punkte]

What to build

The first two screens a user sees: a drag-and-drop upload page that detects the conversation format and shows which participants were found, followed by a processing progress screen that tracks the pipeline stages (parsing → chunking → extracting persona → ready).

Acceptance criteria

• Drag-and-drop file zone accepts .txt files
• After drop: detected format and participant list shown before confirming upload
• User selects which participant to simulate
• Processing screen shows pipeline stages with live status (pending / in-progress / done / error)
• Redirects to persona preview on completion
• Mobile-responsive layout

Blocked by

• Upload endpoint — POST /upload detects format, parses, stores conversation + messages #3