Dialogue · Veo 3 / 3.1 · JSON prompt
two old men at a stone chess table in a city park, autumn leaves on the board, one slides his rook home with two fingers, leans back, and delivers the line; the other freezes mid-sip of coffee.
{
"subject": "two old men at a stone chess table in a city park, autumn leaves on the board",
"action": "one slides his rook home with two fingers, leans back, and delivers the line; the other freezes mid-sip of coffee",
"dialogue": "\"Your king called. He wants a lawyer.\" — dry, without looking up",
"scene": "Washington Square-style park corner, pigeons and passersby blurred",
"shot": {
"composition": "over-the-shoulder favoring the speaker",
"camera_motion": "static, rack focus to the opponent's face",
"lens": "85mm shallow depth of field"
},
"lighting": "soft overcast autumn light",
"style": "warm indie-film realism",
"audio": {
"sfx": "chess piece slide and tap, park ambience, a distant laugh",
"music": ""
},
"color_palette": "autumn amber, stone gray, coffee brown",
"negative_prompt": "text, wrong-size chess pieces, extra fingers"
}Generate with this prompt →The line is the payload ('Your king called. He wants a lawyer.'), so the camera is built to serve it — over-the-shoulder on the speaker, then a rack focus to the opponent's frozen reaction. That focus pull is the comic timing. Keep it warm indie-film realism so the delivery feels lived-in; the negative prompt guards the chess pieces from rendering the wrong size.
Paste the JSON straight into Veo 3 or 3.1, or hit Copy as prose for the flattened sentence version — both work, JSON just keeps the fields easy to tweak. To change the look, re-roll one field at a time (swap the camera_motion or lighting and keep everything else fixed). Want a different subject in the same style? Open it in the Veo 3 prompt generator and edit the fields.
Browse all prompts: Veo 3 library · AI video prompt generator