TokiToki Internal
Atlas + S3
All runs
CompletedComfyUI

ecc6c9f5-51f4-4a33-ba3c-973c2366d09b

Queued
Jul 10, 2026, 12:17 PM
Completed
Jul 10, 2026, 03:06 PM
Execution time
24m 46s
Outputs
1 · 22 MB
Inputs
0
Cached nodes
10

Outputs

1

Click an image to view it full size.

LTX_Director_00094_.mp4

node 37 · 22 MB

Text prompts

3
Positivenode 131 · LTXDirector · local_promptsLTX Director

The sequence runs: The hero sprints forward three steps toward camera — low-angle, water spraying off each impact, camera tracking forward 20% to stay with him. Behind him: the background performers respond in choreographed succession — not simultaneously — each one a beat after the other, each person individual in their movement. The burning motorcycle continues burning — flame turbulence, smoke column rising and bending right with the rain-wind, embers rising. The rain falls through the fire light (orange-tinted drops near the fire), through the police lights (red-tinted), and through ambient night (grey-white). The wet road reflects all of this in real time — the hero's splashing feet sending ripples through the reflected image of the fire and lights. Five seconds in: the camera catches a brief boom-mic level Dutch angle as the hero turns — then self-corrects to the low angle. (Testing: does the model hold camera instruction under motion?) End frame: hero stopped, arm down, looking at camera, rain still falling. Fire still burning. Performers in background, stilled.

2 other text fields— sigmas, filenames, and other non-prompt strings the logger swept up

node 131 · LTXDirector · timeline_data

{"mainTrackEnabled":true,"audioTrackEnabled":true,"motionTrackEnabled":true,"propHeight":90,"globalPropHeight":60,"showFilenames":true,"overrideAudio":false,"inpaint_audio":false,"global_prompt":"","retake_global_prompt":"","retakeMode":false,"retakeStart":24,"retakeLength":48,"retakePrompt":"","retakeStrength":1,"retakeVideo":null,"normalStartFrame":0,"normalDurationFrames":270,"segments":[{"id":"1783685810834kv7cz","start":0,"length":295,"prompt":"The sequence runs:\n\nThe hero sprints forward three steps toward camera — low-angle, \nwater spraying off each impact, camera tracking forward 20% to stay with him.\n\nBehind him: the background performers respond in choreographed succession — \nnot simultaneously — each one a beat after the other, \neach person individual in their movement.\n\nThe burning motorcycle continues burning — flame turbulence, smoke column \nrising and bending right with the rain-wind, embers rising.\n\nThe rain falls through the fire light (orange-tinted drops near the fire), \nthrough the police lights (red-tinted), and through ambient night (grey-white).\n\nThe wet road reflects all of this in real time — \nthe hero's splashing feet sending ripples through the reflected image \nof the fire and lights.\n\nFive seconds in: the camera catches a brief boom-mic level Dutch angle \nas the hero turns — then self-corrects to the low angle. \n(Testing: does the model hold camera instruction under motion?)\n\nEnd frame: hero stopped, arm down, looking at camera, rain still falling.\nFire still burning. Performers in background, stilled.","type":"image","imageFile":"whatdreamscost/`10 Hero_in_rain_action_sequence_202607101739.jpeg","imageB64":"/api/view?filename=%6010%20Hero_in_rain_action_sequence_202607101739.jpeg&type=input&subfolder=whatdreamscost"}],"motionSegments":[],"audioSegments":[]}

node 131 · LTXDirector · resize_method

maintain aspect ratio

Input media

0

This run had no input media — it generated from prompts alone.