Building a 1-Hour AI Concept Album with ChatGPT, DALL·E & Suno — The Hours We Never Counted

Hi everyone — I recently completed my first AI-assisted concept album, and wanted to share the process with this community as a reflection on using OpenAI and other tools for emotionally structured, longform creative work.


:hammer_and_wrench: Tools & Workflow

:mantelpiece_clock: The Hours We Never Counted is a 1-hour cinematic synthwave project composed as 23 interconnected scenes across a Prelude, 4 Acts, and an Epilogue — no vocals, just music, visuals, and narrative flow.

It was built using:

  • :writing_hand: ChatGPT-4o — to develop the narrative structure, pacing, scene titles, emotional tone, and style-of-music prompts used with Suno AI to generate each track
  • :artist_palette: DALL·E (via ChatGPT) — for visual generation, including use of the inpainting feature in DALL·E 3 to refine specific areas — particularly for improving anatomical realism in hands and correcting initially distorted details
  • :musical_score: Suno v4 — for all music generation
  • :film_frames: Wondershare Filmora — for sequencing, visual integration, and post-production

The goal was to explore how AI could support storytelling without words — using sonic and visual cohesion alone.


:headphone: Watch the Project

:play_button: The Hours We Never Counted (YouTube)


:speech_balloon: Thoughts & Questions for Fellow Builders

I’d love to hear how others are experimenting with:

  • Longform AI-generated content — especially in music, video, or storytelling
  • Combining OpenAI with other tools into cohesive creative workflows
  • Pacing and emotional continuity across multi-part AI compositions

How are you approaching narrative or emotional structure in your own AI-powered creative work?


Thanks for building the tools that make experiments like this possible.
I’m looking forward to learning how others are exploring similar ground.