Direct the shot.
Own the light.
Shot Director is a lighting & camera prompt studio for AI stills and film. Browse real demo stills, mix techniques, and copy prompts built for your image — right here on the web.
Explore the libraries
From the still library
How it works
Upload a reference (optional) or skip. Set mood, or pick None to browse freely.
Camera, lighting, anime styles, and combos — each with example stills you can enlarge.
Drop picks into Photo or Film tracks, resolve conflicts, copy or save the prompt.
Paste into your image/video model of choice. Shot Director only builds the language.
Upload an image (optional) and tell it what you're going for, then hit Suggest — you'll jump to the Prompts tab. Prefer to skip the wizard and just browse techniques? Pick None under creative direction, or open 📷 Camera Angles from the sidebar anytime.
1 Upload your image
🖼️
Click to browse or drag & drop an image here
Analyzed in your browser — then adjust brightness, grade, and download a edited still.
No image handy? Skip and answer the questions only →
🎨 Anime Style Library reference
Three views: Mai / Razen still examples, or Character Profile Sheet Prompts — preferred image-only restyle commands (profile-sheet friendly, no video language). Apply to the Photography prompt with → Photo.
📷 Camera Angle Demos Razen reference
One still photo per Camera Shot using Razen (Refrence/man.png) as the model so you can see how framing and angle change. Click a card to enlarge, then apply that camera technique only to your Photo or Film prompt (does not set an anime style).
🎥 Camera Control 3D pose room
Pose articulated figures in a 3D room and orbit the camera to frame your shot. Drag empty space to orbit · scroll to zoom · WASD to fly, Q/E down/up, hold Shift to go faster · pick a joint below (or click a limb in-scene) then drag the ring gizmo to pose it. The detected camera angle feeds straight into your Photo/Film prompt.
❔ How to use Camera Control
Camera
- Drag empty space to orbit · scroll to zoom.
- W A S D — fly forward / left / back / right.
- Q / E — fly down / up. Hold Shift to move faster.
- Keyboard flying is disabled while typing in a field, and while a camera-rail take is playing.
- 🎥 Reset camera (in Camera Rail & Export) snaps back to the default framing.
Posing a figure
- Pick a figure chip at the top, then a Joint from the grouped dropdown — or just click a limb in the 3D view. The selected limb lights up teal so you know what you're about to move.
- Drag the coloured rotation rings to pose that joint.
- Switch Pose limb → Move to drag the whole figure (or a scene object) around instead of a single joint.
- Pose preset applies a ready-made pose to the whole figure in one click; ↺ Reset pose undoes all posing back to the neutral stance.
Scene & objects
- Add object drops a box / platform / pedestal / cube / wall into the scene — drag it with the Move gizmo.
- Every object gets a chip in the Objects row — click one to select it instead of hunting for it in the 3D view, then 🗑 Delete object (or the
Deletekey) removes it. - Figures and objects are rested on the floor automatically — when added, when a pose is applied, and when a scene loads. ⬇ Snap to floor re-rests the active figure by hand: it stands on whatever is directly beneath it, so move a figure over a platform and snap to stand it on top.
- 📸 Screenshot downloads the current view as a PNG. 💾 Save / Load saved store and recall a whole scene (figures, poses, paths, objects, camera) in this browser.
Character paths
- ✏️ Draw path, then click the floor to drop waypoints for the active figure. It walks the smoothed curve through them, turning to face the way it's going, with a real gait cycle.
Escleaves draw mode. - Each figure gets its own coloured path. ▶ Preview walk plays them from wherever the camera is; every path also plays during a camera-rail take, so ⏺ Record captures the camera move and the blocking together.
- A walker steps up onto any platform its path crosses, and every figure snaps back to its original pose when the take ends.
Camera rail (multi-shot moves + video export)
- Build a shot from one or more segments — each has its own move, target (whole cast or one figure), duration and easing, and they play back-to-back in order.
- Custom path: frame a shot, click + Add here to drop a waypoint, repeat — plays a smooth curve through every point you captured, starting from wherever the segment before it ended.
- ▶ Preview plays the sequence without recording. ⏺ Record & download exports a real MP4 (or WebM) video file at your chosen resolution/fps. ⏹ Stop ends the take early — a recording in progress is still saved.
Sending it to your prompt
- → Photo / → Film write the whole scene into that track's prompt: each figure as
image_1,image_2… with its pose, position, spacing/depth from camera, plus the detected camera angle and any camera-rail move. - 📋 copies just the scene-prompt text without switching tabs.
🦾 Pose
🧱 Scene & Objects
🚶 Character Paths
Give the active figure a route to walk. Hit ✏️ Draw path, then click the floor to drop waypoints — the figure walks the smoothed curve through them, turning to face the way it's going, with a real walk cycle. Each figure gets its own coloured path, and every path plays during a camera-rail take, so ⏺ Record & download captures the camera move and the blocking together.
🎬 Camera Rail & Export
Build a shot out of one or more segments — each plays after the last, in order. e.g. segment 1: crash zoom on Figure 1 (2s); segment 2: 360° orbit around the cast (6s, slower).
💡 Lighting Library Master list
Every Lighting Technique (and Lighting + Camera combo) from the Master Shot & Lighting List. Browse, search, enlarge the demo still, or apply a pick straight into your Photo or Film prompt. Use Compare on two cards to view them side by side — same character reference so differences are pure lighting (or lighting + framing for combos).
⚖ Lighting compare
🌱 Shared Seedance Prompts @MusicSatire
Your shared Seedance fight / camera prompts with YouTube Shorts and X previews. Open the video or copy the prompt. Channel: youtube.com/@MusicSatire · X: @MusicSatire
🏆 Community Project of the Week
Made something with Shot Director? Share your prompt and a picture of the result — the best get featured as Project of the Week.
Submit your prompt & image
🛡️ SFW only. Every submission is human-reviewed before it appears — nothing goes live automatically. No nudity, hate, or abuse.
📐 Camera Angles Guide
🌱 Seedance 2.5 Guide
🎬 Skyebrows Prompt Library reference
Each original preset has a still keyframe demo. Newer entries scraped from @Skyebrows (full Grok text when posted on X + Grok Imagine links). Click a thumbnail to enlarge, apply with → Photo / → Film, or pick from a prompt's 🎬 Skyebrows preset dropdown.
All presets (dropdown + stills)
🪄 Transparent Background Tool
Upload an image with a solid green, blue, or white background and knock it out to transparency, then download a PNG. Processing happens in your browser on this page.
🖼️
Click to browse or drop an image
…or click the preview to sample the exact colour. Current target: —
How close a colour must be to the target to be removed.
Colour ramp: how gradually the cut fades as pixels approach the target colour.
Softens the cut-out shape itself — best for wispy hair and fine lines.
Neutralises leftover green/blue colour cast on the kept edges (no effect on a white background).
Upload an image to preview the cut-out here.
👾 Sprite Maker
Turn an uploaded image into a pixel-art sprite — pixelate it, snap it to a retro palette, knock out the background and add a sprite outline, then download a transparent PNG. This is an in-browser pixel-art converter (not an AI redraw); for a full AI "reimagine as a game character", build a prompt on the Prompts tab and run it through an image model.
🖼️
Click to browse or drop an image
Longest side of the sprite grid — smaller = chunkier pixels.
Upload an image to make a sprite.
🧝 AI Character Sprite
Upload a character image and turn it into an LPC-style pixel-art sprite with a generative AI. Claude can't make images, so this needs a backend: a free local Stable Diffusion server (AUTOMATIC1111 / Forge / ComfyUI — best, because it works image-to-image off your upload) or a cloud image API (OpenAI, Grok/xAI, Stability…) with your own key. The prompt below always works — copy it into any AI tool even without wiring up Live Generate.
🖼️
Click to browse or drop a character image
⚡ Live Generate (advanced — needs your model/key)
Higher = more stylised / less like the original photo.
Upload a character image, copy the prompt into your AI tool, or run Live Generate.