AI Video Automation Pipeline

I want to take a raw idea and have it emerge on Facebook, Instagram, and YouTube as a polished, two-minute video—all without my manual intervention. The agent you build must handle three big jobs for me:

• Content creation. From a topic that I supply—or one the agent dreams up itself—it should draft a script, add a natural-sounding AI voice, select fitting stock or generated visuals, lay in background music and on-screen captions, and package everything into a cohesive edit.
• Customisable publishing. Flexible scheduling, post formats, and asset-level tweaks are essential because publishing options are my top priority. I need to review the finished video, approve it, then watch the agent push it live (or schedule it) on the three platforms listed above. Dynamic title, description, and hashtag generation must be part of this flow so every post is fully optimised.
• Scalability. The architecture should make it easy to move from occasional uploads to daily automated releases without a full rebuild.

I am open to whatever blend of n8n, Make, bespoke scripts, AI video APIs (Runway, Pika, or similar), voice services (e.g., ElevenLabs), and platform APIs you feel is best. In your reply, briefly outline:

– The overall workflow you would stitch together and why
– Specific tools or APIs you’d pair for scriptwriting, voice, video assembly, and multi-platform posting
– A timeline with milestone estimates and rough cost breakdown (build plus ongoing API/tool fees)
– One or two examples of similar automations or AI-driven video projects you have delivered

A generic copy-paste proposal will be ignored; just give me a concise plan that proves you know how to stitch these parts into a single dependable system.