AI is becoming a native device and operating-system layer
A targeted GitHub cohort and the existing local-runtime archive measure browser, mobile, desktop, and edge deployment surfaces.
Supported by the targeted cohort
405 relevant projects and models span desktop (51.1% of the cohort), mobile (20.2%), edge (49.6%), and browser (28.9%) surfaces. Device AI is broadening beyond a local-desktop niche, although deployment remains desktop-heavy.
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 268217 loaded
Feb 2610748 loaded
Mar 2619775 loaded
Apr 2617075 loaded
May 2615574 loaded
Jun 2618599 loaded
Jul 266517 loaded
02
What exists inside the category
One item may contain more than one feature.
Browser runtimes
117 · 28.9%
Mobile deployment
82 · 20.2%
Desktop deployment
207 · 51.1%
Edge and embedded
201 · 49.6%
Offline and private
68 · 16.8%
03
Representative projects
github-local-runtimejamiepine/voicebox40,620 starsgithub-local-runtimeAlexsJones/llmfit29,284 starsgithub-local-runtimejundot/omlx17,754 starsgithub-local-runtimeopen-pencil/open-pencil6,823 starsgithub-local-runtimeraullenchai/Rapid-MLX3,274 starsgithub-local-runtimenicedreamzapp/claude-code-local2,944 starsgithub-local-runtimefikrikarim/parlor1,894 starsgithub-local-runtimetrevin-creator/autoresearch-mlx1,732 starsgithubTypeWhisper/typewhisper-mac1,617 starsgithubapple/coreai-models1,431 stars
04
Cross-source validation
These publications are not part of the primary numeric cohort.