Aesthetic-first AI image generator with style control and creative direction tools
AI rap video generator that turns two photos into a personalized rap duet
Stage
Not given
Not given
Founded
Not given
Not given
Based in
Not given
Not given
Pricing model
Not given
Paid
Pricing
Annual plans (50% off launch offer): Starter $9.9/month (9,600 credits yearly), Pro $19.9/month (24,000 credits yearly), Studio $44.9/month (72,000 credits yearly). Monthly and one-time credit packs also available.
One video $6.99, three-video Duo pass $14.99 ($5.00 per video), ten-video Crew pass $39.99 ($4.00 per video). Passes are one-time purchases with no subscription and never expire.
Key features
Expressive text-to-image generation with aesthetic diversity
Style references with adjustable per-reference strength
Moodboard direction for coherent visual language
Adjustable creativity levels from raw to high
Medium and Large model variants
Krea 2 Turbo speed-optimized variant
Up to 9 reference image slots per generation
4K resolution export capability
Prompt expansion for richer visual briefs
Commercial-use licensing on all paid plans
Scene store with sets and scenes
Two separate photo uploads (one face per performer)
Original lyrics written for the scene
AI-generated rapped vocals and beat
Mouth movement lip-sync
Vertical 9:16 MP4 video format
Watermark-free downloads
Renders in 1–3 minutes
Browser-based, no app installation
What makes it different
Separates subject intent from aesthetic intent for precise style exploration
Top-ten placement on Artificial Analysis text-to-image leaderboard
Open-source model weights available on Hugging Face
Wider aesthetic vocabulary covering raw, photographic, illustrative, and painterly styles
Designed for early art direction when exploration matters more than a single polished answer
Scene store catalog with multiple sets and scenes, expanded weekly
Two separate photos required allows friends in different cities to collaborate
Reuse same two photos across any scene in the catalog
Continuous single-shot performance, not a slideshow