ModelHub provides model-installation entry points, the LLM plugin exposes JNI decoding, and speech modules provide recording, recognition and synthesis paths.
AI engineering / Native integration
On-device AI · Model, inference and speech plugins
A mobile plugin stack for model resources, native inference and audio capture/synthesis, centered on resource lifecycle and local/server routing.
WORK & CONTRIBUTIONS
My work on this project.
Separates web calls, Android bridges, model resources and native lifecycle so application logic can choose inference paths and stop/release session resources.
The reviewed app restricts local mode to Android. LLM and speech iOS entries remain echo placeholders, so parity is not claimed.
NEXT PROJECT