God's Way Foundation@godswayfoundinc on XHire God'sView original post

Stop describing real locations to AI. Give it coordinates. 👇 Every Times Square shot in 'Taxi Ride From Hell' started as one pin in Google Maps: "40.75726, -73.98601" ... that's it. Here’s the workflow: 1⃣ Drop a pin in Google Maps exactly where your camera would stand. Right-click it and copy the coordinates. 2⃣ Paste the coordinates BY ITSELF into your image prompt. Add the time of day, camera height and era. 3⃣ Use that image as the location reference in your video prompt, then direct your characters inside it. “Busy New York street” gets you a generic city. Coordinates get you the real buildings, the real sightlines and the real street layout, and the location holds from shot to shot. 🚨 Pause on the split screen in the video. Same tkts booth, same flag billboard, same crosswalk. Full film in the replies. Made on @itsPolloAI

A tutorial turning a Google Maps pin into coordinates, reference image and an AI-generated Times Square shot.

435 views
14 likes
3 reposts
4 replies
5 bookmarks

Posted Oct 26, 20260:27Portrait · 1080×1350

Short films Tutorials & breakdowns

Tutorial Mixed styles Cities & architecture Tech & gadgets

This video is God's Way Foundation’s work, shown from their public post. See it on X.

More from God's Way Foundation

All 12 videos

1:29

TAXI RIDE FROM HELL 🚕🔥 He’s late for the 2:39 at Grand Central. His Rasta cabbie has rockets. Times Square, the Empire State, and 1999 NYC from the sky. My entry for "One Minute in New York", made with @itsPolloAI 🎬

502X

1:00

Before the group chat, there was the block. A little of how Brooklyn used to feel. Summer in ’90s Bed-Stuy; hydrants running, plates passing, and everybody outside. An little tribute to the cookout. Made with Seedance 2.5 🎬 Song🎵 “People Make the World Go Round” — The Stylistics

450X

2:27

“SHOW, DON’T TELL!” 🎬 AI Filmmakers: Your world had a history before your main character showed up. How do its people tell it? For this scene in "The Alabaster Jar", we built a street performance inside Ashret. Fire breathing, dancing, drums, actors playing up to the crowd, and a storyteller passin

568X

0:301:1

Stop putting your AI characters straight into scenes 🫵😠👎 Here's what you do: give them a screen test first! 🎬 And here's why this is useful: the model reads your character's face and builds their voice from it, even when you tell it the accent...but let's go in depth: 🗣️ THE MODEL CASTS THE VOICE

3.9KX

1:28

🎥 Steal Hollywood's oldest trick so that Seedance 2.5 can hide its mistakes in a fight. Here's the camera setup that serves as the FIX 👇 Hands, feet, contact... the model still gets some of it wrong. The trick is making sure the viewer never gets a clean look at it. This fight is three Seedance 2.

3.3KX

1:59

Your Seedance 2.5 dialogue is too clean🧹 Try inputing dialogue in your prompts that reflect natural language. SImilar to shaky handheld camera shots, it's organic enough to come across as authentic. Three things I do about it in my prompts: 1️⃣ I write stutters and in natural language. Not "east,

11.4KX

0:14

Stop placing everything at the center of your frame with Seedance 2.5 🫵😠👎 Short answer: It doesn't feel like film...but let's go in depth: 📐 WHERE EVERYTHING SITS: In the key frame of this clip, my character is in the LEFT THIRD. The sign he's looking at is in the RIGHT THIRD, high. Neither one

1.7KX

0:30

👇I found the one word that fixed my audio popping and jarring cuts between Seedance 2.5 clips. The word "beat" is your best friend when you're pacing a video. Type it into a prompt and the model gives you roughly a second of nothing. No dialogue, no action, just the shot continuing to breathe. I u

1.8KX