AI creative tools
360° Orbit Video
Turn one photo into a full 360° camera orbit that lands back on its first frame: MiniMax H3 with the 360 Orbit LoRA keeps the scene frozen while only the camera moves.
MiniMax H3 + Orbit LoRA · first = last frame · 3 s loop
3D POSE POSERAnime Pose Control
Pose a 3D mannequin with drag-to-pose IK, then redraw it as a witch, mage or vampire in your exact pose.
Pose poser · 3D mesh control · from $0.08
MULTI-CHARACTER SCENESAnime Scene Control
Place several characters with region maps plus 3D poses, and describe each one in the prompt.
Regional prompts + 3D pose · from $0.10
DEPTH CONTROLAnime Depth Control
Keep the exact volume and layout of a photo or render while redrawing it as anime art.
Depth map control · from $0.10
SKETCH TO ARTAnime Sketch Control
Turn a rough sketch or line art into a finished anime illustration that keeps your lines.
Sketch / edge control · from $0.12Character Recast
Replace the people in a video with new people from reference photos: H3 Max tracks every person across shots and keeps the motion, camera, cuts and soundtrack.
MiniMax H3 Max recast · up to 4 people · 30 s · original audioCharacter Swap Music Video
Redraw the first frame with new characters, then H3 re-performs every shot with the same moves on the original soundtrack.
GPT Image 2 frame · MiniMax H3 reference-to-video · original audioExact Motion Character Swap
Replace each performer inside the original footage with pose-exact motion transfer; cuts, camera and soundtrack untouched.
GPT Image 2 frame · Wan 2.2 Animate per performer · YOLO pose compositeCharacter Swap LoRA
Replace the people in a video with new characters: a character-swap LoRA on MiniMax H3 keeps the room, camera and motion, and the original soundtrack goes back on.
RA2 character frame · MiniMax H3 + swap LoRA · original audioReference Video Studio
Direct a video from reference images, a motion video and a soundtrack; follows the choreography and cuts, up to 60 seconds.
Seedance 2.0 Mini / Pro · up to 60 s · soundtrack kept
CHARACTER ARTAnima Art Studio
Anime-native character design and polished illustration.
Anima-2.9B · seed 18467291
BULK TEXT TO IMAGEMake Image
Generate many variants at once with RA2, one prompt per line.
RA2 · multi-variant · $0.04 per image
CAMERA CONTROLCinematic Cameras
Image generation steered by dolly, crane, macro, and lens-direction presets.
RA2 · GPT Image 2 · camera presets
COMPOSITION-AWARE RESIZESmart Resize
Adapt one image to stories, posts and banners. AI recomposes the scene for exact dimensions instead of stretching it.
fal Smart Resize - exact sizes - up to 6 formats
TRANSPARENT CUTOUTSImage Background Remover
Remove backgrounds from people, products and characters without redrawing the subject. Download a transparent PNG.
BiRefNet - transparent PNG - $0.03 per image
LIGHTING CONTROLRelight
Move the key light. Adjust direction and mood of any photo with IC-Light v2.
fal IC-Light v2 · directional relightingMASKED EDITINGInpaint
Brush an area, describe the change—everything outside your mask stays pixel-identical.
GPT Image 2 edit · feathered mask composite
ENHANCEImage Upscale
2x creative upscaling that recovers texture and detail without plastic artifacts.
fal creative-upscaler · 2xOUTPAINTINGExtend Image
Grow the canvas in any direction or zoom out while the scene keeps extending.
extend-image · expand or zoom out
REFERENCESSoul Moodboard
Drop 2–6 references and fuse them into one focused visual direction.
Nano Banana 2 · GPT Image 2 · multi-reference
IMAGE MODELNano Banana 2
Google Gemini-flash image generation and reference editing at speed.
or/gemini-3.1-flash-image via OpenPaths
IMAGE MODELGrok Imagine 2.0
High-resolution image generation by xAI with 2K quality tier.
grok-imagine-image · 1K/2KIMAGE MODELFLUX.2
Speed-optimized detail from Klein, Dev, and Pro lanes.
FLUX.2 Klein · Dev · ProIMAGE MODELGPT Image 2
4K-capable images with near-perfect text rendering.
gpt-image-2 · always metered
IMAGE EDITINGStyle Transfer
Upload a source image, describe a new visual world, and keep the composition while the edit runs.
OpenPaths image-edit · GPT Image 2 route
TEXT TO IMAGEH3 Image
High-detail finished frames from the H3 visual world model.
MiniMax H3 · real generationREFERENCE EDITH3 Image Editor
Regenerate a source while preserving identity and geometry.
H3 REF2VA · real editCharacter Animator
Transfer body movement, expression, and timing from a driving clip.
Wan Animate 2 · real 5s Standard outputWan Animate Move
Make a complete reference image follow a driving performance.
Wan 2.2 Animate · Move · 480p–720pWan Animate Replace
Swap the performer while keeping the source video scene and lighting.
Wan 2.2 Animate · Replace · 480p–720pPose Video Transfer
Turn a driving performance into a witch, mage, creature, or completely new character.
Pose motion control · ordinary video or prepared passDepth Video Control
Rebuild a clip while preserving moving depth, camera path, and spatial volume.
Depth control · ordinary video or prepared passCanny Video Control
Keep silhouettes and motion locked while redrawing every frame in a new world.
Canny edge control · ordinary video or prepared passHED Video Control
Preserve soft organic contours for expressive creature, face, and illustration transfer.
HED contour control · ordinary video or prepared passMLSD Video Control
Hold straight lines and perspective stable while transforming designed spaces.
MLSD line control · ordinary video or prepared passVideo Background Remover
Create a transparent foreground while retaining original pixels.
RVM · real transparent outputVideo Remake Agent
Detect every shot, guide a coherent restyle, and preserve the exact timeline and soundtrack.
DeepSeek vision · GPT Image / Nano Banana · H3 Max motionCinematic Trailer Agent
Sketch characters, lock identity sheets, repair drift, then animate accepted start frames with Gemini speech, a score, and H3.
RA2 sketch · GPT Image 2 · Gemini TTS · H3 · Manifold music
MUSICAI Music Generator
Describe a song; an LLM writes the arrangement and lyrics, then the Manifold music model records a full 48 kHz stereo track.
LLM composer · Manifold music model · 48 kHz stereoLofi Loop Studio
Turn any track into a seamless looping lofi video with generated cover art and an audio visualizer.
RA2 cover · parallax loop · 6 visualizers
GOOGLE MUSICLyria 3 Music Studio
Direct full songs, scores, vocal hooks, and loops with timed structure and compact Opus output.
Lyria 3 Pro · Clip · OpenPaths · 44.1 kHz stereo
MULTI-SPEAKER TTSGemini Voice Director
Stage expressive narration and two-person scenes with 30 voices, accents, pace, and emotion cues.
Gemini 3.1 Flash TTS · OpenPaths · 30 voices
TEXT TO SPEECHVoice Studio
Expressive narration, dialogue, and audio scenes from a script across seven voice models.
Seed Audio · ElevenLabs v3 · Gemini VoiceVIDEO GENERATORS
KlingVeoSeedanceWanLTXManifoldHappy HorseRA2VAll