wmdrift
Diagnostics for autoregressive video world models - quantifies camera drift, temporal jitter, memory attenuation, and quality decay from rollouts.
Details
- External ID
- 1367932755
- Source
- GITHUB
- Company
- —
- Product
- wmdrift
- Website domain
- github.com
- Launched
- Sept. 13, 2026
- Cohort
- —
- Upvotes
- 16
- Upvotes percentile
- 0.5194722008711248
- Tags
- clip, computer-vision, diagnostics, evaluation, lpips, opencv, pytorch, video-generation, visual-odometry, world-models
- Fetched at
- Sept. 17, 2026, 5:02 p.m.
- Updated at
- Sept. 17, 2026, 5:02 p.m.
Enrichment
- Theme
- ai video generation and editing tools
- Vertical
- Horizontal
- Function
- Observability & eval
- Audience
- Developer
- AI stance
- AI-native
- Project type
- Hobby / open-source project
- Normalized one-liner
- diagnostic evaluation toolkit for video world models
- Manually corrected
- False
Could you build this?
Partial The CLI harness and reporting metrics can be vibe-coded, but calculating precise camera drift, temporal jitter, and rollout decay across autoregressive video world models requires specialized computer vision and optical flow evaluation pipelines.
What it would actually take: Requires integrating computer vision toolkits (RAFT/FlowNet for optical flow, Structure from Motion / COLMAP for camera pose estimation, and perceptual quality metrics like FVD/LPIPS). The primary difficulty is robustly separating camera motion drift from semantic object deformation across autoregressive rollouts without ground-truth 3D tracks. Requires domain knowledge in generative video architectures and 3D vision geometry.
Competitors
Other products that read as similar to this one — 765 launches clear the similarity bar, closest 8 shown.
Attention rank: #375 of 766 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 317 days after the earliest competitor.
- VideoMetricsLab · github · 2026-09-21 · 12 upvotes · similarity 0.60
- Minimax-H3-Latent-Continuation · github · 2026-09-18 · 31 upvotes · similarity 0.59
- PixelUp · hn · 2026-07-18 · 6 upvotes · similarity 0.57
- gemos-still · github · 2026-09-12 · 14 upvotes · similarity 0.57
- 30min video analysis for $0.003 via frame-tiling and Vision API · hn · 2026-01-18 · 7 upvotes · similarity 0.57
- glance-speedlab · github · 2026-09-23 · 48 upvotes · similarity 0.57
- Crow-DLSS5-Video-Image-Converte · github · 2026-09-11 · 7 upvotes · similarity 0.56
- VideoPhysEdit · github · 2026-09-26 · 9 upvotes · similarity 0.55
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a observability & eval tool for Media & entertainment yet.