Generative media is becoming a controllable production system
GitHub and Hugging Face cohorts measure control features relative to the broader generative-video supply.
Supported by the targeted cohort
Control is becoming part of the product surface: editing appears in 8.6% of the cohort and production workflows in 8%. Identity, audio synchronization, and explicit keyframe control remain much thinner, so the production stack is forming unevenly.
01
Search-match dynamics
Bars show monthly GitHub total_count plus captured Hugging Face models; the right column shows downloaded items used for detailed analysis.
Jan 2643725 loaded
Feb 2647829 loaded
Mar 2669235 loaded
Apr 2653217 loaded
May 26915486 loaded
Jun 26817450 loaded
Jul 26321152 loaded
02
What exists inside the category
One item may contain more than one feature.
Keyframe and motion control
5 · 0.4%
Identity consistency
5 · 0.4%
Iterative editing
103 · 8.6%
Synchronized audio
73 · 6.1%
Production workflows
96 · 8%
03
Representative projects
githubcalesthio/OpenMontage41,050 starsgithubHBAI-Ltd/Toonflow-app11,929 starsgithubForget-C/Jellyfish5,581 starsgithubFireRedTeam/FireRed-OpenStoryline3,133 starsgithub-archiveluoluoluo22/jianying-editor-skill2,403 starsgithubPKU-YuanGroup/Helios1,995 starsgithub-archiveLingyiChen-AI/AIComicBuilder1,642 starsgithubbytedance/Bernini1,161 starsgithub-archivedramaclaw/dramaclaw1,054 starsgithubPurpleDoubleD/locally-uncensored942 stars
04
Cross-source validation
These publications are not part of the primary numeric cohort.