Generate a start frame with ChatGPT, then have HeyGen make a ghost crawl out of an X feed
toolstelegraph · x · 2026-10-31
A hands-on image-to-video tutorial: first use ChatGPT with a detailed prompt to generate a smartphone screenshot of an X (Twitter) dark-mode feed as the start frame, then feed that image plus a reference creature photo into HeyGen (Claude integration also works) with a prompt.
Key prompt details: locked-off static shot with no camera movement or cuts, both posts stay visible throughout; the creature is a tall gaunt woman with long wet black hair, pale gray-green skin and a jagged-tooth grin, matched exactly to reference image 2; the core effect is a 3D "out of the frame" pop-out where she emerges from the news image while the X feed layout stays intact.
More from Multimodal
- AI Video Recreates Cinderella-With-Sword Stairwell in Hand-Drawn Style — Pikturlab · 2026-10-31
- AI Video Nails Retro Sitcom Look With a 1950s-Style Robot Hug — Darri3D · 2026-10-31
- After 2 Years of Chasing Perfect Lipsync, He Nails It with a Magnific Workflow — techhalla · 2026-10-31
- Mutuals open-sources character harness: voice chat avatars with reactive faces and bodies — gdechichi · 2026-10-31
- ComfyUI-Dials: prompt pairs become 14KB sliders controlling video motion & mood, no LoRA training — MatlowAI · 2026-10-31
- One Midjourney sref Code, Three Consistent Noir Animal Characters — sergeantsref · 2026-10-31