English quick take
The local lab applies a preview grade and exports a short silent clip. The prompt recipes show how to ask an image-to-video model for the same effect without turning the footage into mush.
Girigo AI Studio 9 min read
Give a short clip a deliberate horror texture: analog scanlines, a cold grade, a fog layer, or one controlled glitch beat.
English quick take
The local lab applies a preview grade and exports a short silent clip. The prompt recipes show how to ask an image-to-video model for the same effect without turning the footage into mush.
Local clip lab
Choose a clip you own. The browser trims and grades a preview; no file is uploaded.
Choose a short clip to start. Nothing leaves this browser.
Prompt studio
Pick a look, set the direction, and copy a prompt that names what must stay fixed.
Direction: cinematic portrait; preserve identity; one person.
Use the provided image as the exact starting frame. Create a five-second continuous analog-horror shot. Keep the subject's face, clothing, and body proportions consistent with the reference. Add subtle VHS scanlines, a faint horizontal tracking wobble, soft chroma bleed, and a cold fluorescent grade. The subject blinks once and slowly looks toward the camera. Keep the camera mostly steady with a tiny handheld drift. Maintain realistic motion and stable anatomy. Do not add a second person, do not add a jump scare, do not cut the scene, and do not add readable text or watermarks.
Three useful outcomes
Scanlines, chroma bleed, grain, and a cold grade do most of the work.
Open the studioTrim a short clip and export a silent preview in the browser.
Open the studioA single glitch beat is easier to control than a cascade of effects.
Open the studioPrompt lab
A budget route is enough for a texture-first test.
Use the provided image as the exact starting frame. Create a five-second continuous analog-horror shot. Keep the subject's face, clothing, and body proportions consistent with the reference. Add subtle VHS scanlines, a faint horizontal tracking wobble, soft chroma bleed, and a cold fluorescent grade. The subject blinks once and slowly looks toward the camera. Keep the camera mostly steady with a tiny handheld drift. Maintain realistic motion and stable anatomy. Do not add a second person, do not add a jump scare, do not cut the scene, and do not add readable text or watermarks.
Avoid: second person, jump scare, scene cut, readable text, watermark
Fast queues make it easier to compare glitch timing.
Animate the provided still as a five-second basement horror shot. Preserve the subject's identity, pose, clothing, and lighting. Add one controlled glitch beat at the two-second mark: a brief frame tear and a single flicker of the practical light, then return to the original image. Keep the camera steady. The subject does not move except for a slow breath. Maintain realistic anatomy and consistent color. Do not add extra people, a monster, a scene cut, or text and watermarks.
Avoid: extra people, monster, scene cut, text, watermark
Use a stronger motion route when the light itself must move.
Use the provided image as the starting composition. Create a five-second atmospheric horror clip. Keep the subject's face, costume, and proportions unchanged. A single warm lantern swings gently in the foreground, casting moving shadows across the wall. Fog drifts low across the floor. The camera performs a very slow lateral drift. The subject holds still and blinks once. Maintain realistic motion, stable anatomy, and consistent lighting. Do not change the person, add a second person, introduce a jump scare, cut the scene, or add text and watermarks.
Avoid: face change, second person, jump scare, scene cut, text, watermark
WaveSpeed, SeeAPI, fal, Replicate, Kie
| Provider | Best for | Typical route | Docs |
|---|---|---|---|
| WaveSpeedBudget-first | Low-cost image edits and image-to-video drafts | Seedream edit, FLUX Kontext, Wan image-to-video | Docs |
| falSpeed-first | Fast hosted queues and strong model coverage | FLUX, Kling, and other hosted media models | Docs |
| KieCharacter-first | Nano Banana image edits and Kling video routes | Nano Banana, Kling video | Docs |
| SeeAPIRouting layer | One API surface for several model families and fallback routes | Image and video inference endpoints | Docs |
| ReplicateExperiment-first | Broad catalogue for experiments and batch comparisons | Community and first-party image/video models | Docs |
Step by step
Failure modes
Editorial brief
Horror video effects are a post-production problem disguised as a generation problem. A model can add motion, but it is not the best tool for consistent scanlines, a stable timecode, or a clean color grade. The reliable workflow is to generate a simple motion plate, then add the texture in the browser or an editor where you can control it frame by frame.
Use the local lab for the parts you can see immediately: crop, contrast, color temperature, vignette, and grain. Use the AI prompt only for motion that would be expensive to shoot. If the model adds a glitch you like, export it and treat it as a texture plate. Do not ask the model to reproduce the same glitch exactly; ask for a clean motion beat and add the repeatable effect yourself.
Deeper guide
The prompt is only half the job. The field note walks through the failure modes, the identity check, and the model comparison behind this page.
Open the field noteFAQ
Yes. The local lab applies a preview grade, scanline overlay, and vignette in the browser. Use the AI prompts when you want the model to generate new motion, not just a filter.
The browser lab exports a visual preview. Audio is intentionally left out so the export works consistently. Add sound in your editor after the picture is locked.
Texture is usually better handled in post than in a video model. Use a cheap image-to-video route for motion and add scanlines, grain, and chroma bleed in the local lab or an editor.
The browser lab exports WebM where MediaRecorder is supported. Convert to MP4 with your editor or ffmpeg if the platform requires it.
Checked sources