Build the full script generation pipeline using LLMs. Accepts user input + context, generates structured script output, and stores versioned content for reuse or editing.

Scope:

• Build backend service to invoke LLM with contextual prompts
• Create script-generation endpoint (POST /api/scripts)
• Connect UI to trigger generation (based on user input + config)
• Add metadata tagging (tone, type, tags, etc.)
• Start handling streaming or async generation if needed
• Version generated scripts (for later edits or history)