How to Make H3 Video from Text in 2026

Write a shot, not a paragraph: subject, action, camera, light, native audio. A Tideglass harbour clip with the exact prompt.

Try Netwrck free Open the tool

H3 is the model behind Netwrck text-to-video. It renders picture and diegetic sound in one pass, so the prompt has to direct both. A mood board in prose gives pretty fog. A shot list in prose gives a camera move you can cut.

Generated on Netwrck

Prompt

PHOTOREAL CINEMATIC. Tideglass harbour at blue hour, brass tide-clock towers, indigo fog.
ACTION 0-5s: slow crane over wet cobbles as lanterns bloom and a fishing skiff slides through the fog.
AUDIO: distant bells, water slap, gulls. No music, no dialogue, no text.

Tideglass harbour still generated for this series

H3 image-to-video of the harbour, native audio · 5s

H3 text-to-video grammar on a different subject · 5s

Five seconds. Preview to iterate, Native to publish.

Copyable prompt

PHOTOREAL CINEMATIC. Tideglass harbour at blue hour, brass tide-clock towers, indigo fog.
ACTION 0-5s: slow crane over wet cobbles as lanterns bloom and a fishing skiff slides through the fog.
AUDIO: distant bells, water slap, gulls. No music, no dialogue, no text.