Open-source AI model for high-quality video generation
Convert video clips into per-frame depth map videos with AI estimation
Stage
Not given
Startup
Founded
Not given
2026
Based in
Not given
Not given
Pricing model
Open source
Freemium
Pricing
Not given
Credits consumed per second of input video: VDA-Small 2 credits/second, VDA-Base 3 credits/second, VDA-Large 4 credits/second, with a 2-credit minimum. Subscription plans available: Basic $6.90/month with 100 credits, Pro $13.90/month with 250 credits, Max $34.90/month with 750 credits, Pro Max $59.
Key features
Generates videos from text prompts
High-fidelity motion at 30 frames per second
10 billion parameter architecture
480p resolution with future HD support
Videos up to 5.4 seconds in duration
API integration available
Open source under Apache 2.0 license
Three model tiers: VDA-Small for quick previews, VDA-Base for balanced speed and detail, VDA-Large for finest detail
Grayscale and Turbo visualization modes
Upload formats: MP4, MOV, WebM, M4V, GIF up to 60 seconds and 200MB
Per-frame temporal stability with frame-by-frame alignment
Depth output as MP4 ready for compositing software
No per-frame rotoscoping required
Real-time duration verification for accurate pricing
Relative depth output, not calibrated physical distance
Support for depth-driven visual effects like fog, defocus, and parallax
What makes it different
Open-source AI model freely available under Apache 2.0 license
Exceptional alignment with textual prompts
Temporally stable depth passes without depth cameras or rotoscoping
Three model tiers with transparent per-second credit pricing
Instant price confirmation before processing any video
Dual visualization modes for different use cases
Works with everyday video formats and up to 60-second clips