Mindvalley Stories
Case Study · AI Mastery

Two custom GPTs became the backbone of a full AI video production workflow

A Video creator GPT and a Songwriter GPT guided scenes, lyrics and AI clips into one finished video blending live and AI footage.

This creator did not just try AI tools one at a time. They built their own custom GPTs to plan scenes and write lyrics, then pulled clips, music and live footage together into one video. It is a clear picture of what becomes possible when you learn to direct AI rather than simply use it.

Pavla Nygaard
A video creator who builds custom AI assistants to plan and produce music videos.

Custom GPTs for AI video production

Before

Making a video usually means juggling many jobs at once: planning scenes, writing music, shooting footage and editing it all together. The submission does not describe their starting point in detail, but the challenge was clear. They needed a way to bring many separate tools into a single, ordered process.

“I created a Video creator customGPT for scene identification.”

In their own words

The turning point

During the program they began building their own custom GPTs. They created a Video creator GPT for scene identification and a Songwriter GPT for lyrics. Instead of starting from a blank page, they now had two assistants shaped for the exact jobs they needed done.

“I first worked with the Video creator to plot out the sequence of the scenes, then I went to RunwayML and SORA to create AI clips.”

In their own words

After

They now run a step-by-step production workflow. They plot the sequence of scenes with the Video creator, generate AI clips in Runway ML and SORA, and create a song with lyrics relevant to the video using Suno. Finally they splice live footage, AI clips and audio together in CapCut.

What it lets people do

The workflow shows how one person can plan, score and assemble a video with AI assistants handling each stage. Custom GPTs keep the scene planning and songwriting focused, so the creative choices stay with the maker.

What they learned

The lesson they take from it is that the real power lies in orchestration. Each tool does one job well, and building your own assistants turns a scattered set of apps into a single creative process that mixes AI with real, live footage.

“I spliced the combination of live and AI video clips and audio in CapCut.”

In their own words

Custom GPTs for AI video production was built during AI Mastery.

Read the full submission
Tools Used: ChatGPT, DALL-E, Suno, Runway ML, SORA. CapCut (not Al, for video editing). Live videos.  Methodology: I created a Video creator customGPT for scene identification. I also created a custom Songwriter GPT for lyrics. I first worked with the Video creator to plot out the sequence of the scenes, then I went to RunwayML and SORA to create AI clips. I created a song with lyrics relevant to the video. I spliced the combination of live and AI video clips and audio in CapCut. Tools Used: ChatGPT, DALL-E, Suno, Runway ML, SORA. CapCut (not Al, for video editing). Live videos.  Methodology: I created a Video creator customGPT for scene identification. I also created a custom Songwriter GPT for lyrics. I first worked with the Video creator to plot out the sequence of the scenes, then I went to RunwayML and SORA to create AI clips. I created a song with lyrics relevant to the video. I spliced the combination of live and AI video clips and audio in CapCut.

Written 22 January 2025 Submitted by the storyteller

Every story is written by a Mindvalley student or by someone our teachers have worked with. How we collect and check stories

More from AI Mastery

Custom GPTs for AI video production: a AI Mastery case study | Mindvalley Stories