GenieX
Run frontier LLMs and VLMs locally on Qualcomm devices across NPU, GPU, and CPU with a few lines of code
What is GenieX?
Run frontier LLMs and VLMs locally on Qualcomm devices across NPU, GPU, and CPU with a few lines of code
Overview of GenieX
What it is
GenieX is an on-device Gen AI inference runtime for Qualcomm devices. It allows users to run frontier LLMs and VLMs locally on the Hexagon NPU, Adreno GPU, or CPU with a few lines of code. The runtime is exposed through a CLI, Python, Kotlin/Java, Docker, and an OpenAI-compatible server, all built on a single C SDK. It is the community version of Qualcomm GENIE.
Key features
- Supports GGUF models from Hugging Face, ModelScope, or Docker Hub, as well as pre-compiled bundles from Qualcomm AI Hub.
- Runs on Qualcomm Snapdragon platforms including Windows ARM64, Android, and Linux ARM64.
- Provides two runtimes: llama.cpp (for GGUF models on NPU, GPU, or CPU) and Qualcomm AI Engine Direct (for pre-compiled bundles on NPU only).
- Offers multiple interfaces: CLI, Python (with a transformers-like API), Kotlin/Java SDK, Docker, and an OpenAI-compatible server.
- Includes a model list command to browse available Qualcomm AI Hub bundles for the user's chipset.
Use cases
GenieX is suitable for developers and researchers who need to run large language models and vision-language models locally on Qualcomm devices. Use cases include on-device AI inference for mobile apps, edge computing on IoT devices, and prototyping with local LLMs without cloud dependencies.
Getting started
See the repository README for detailed installation and quickstart instructions. Example commands from the README include:# Install CLI on Linux ARM64 (no sudo):curl -fsSL https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-geniex/install.sh | sh# Run a GGUF model from Hugging Face:geniex infer google/gemma-4-E4B-it-qat-q4_0-gguf# Install Python package:pip install geniex# Pull and serve a model with OpenAI-compatible API:geniex pull ai-hub-models/Qwen3-4B-Instruct-2507geniex serve
License and pricing
GenieX is open source under the BSD 3-Clause license. Use of this project is also subject to Qualcomm's Terms of Use.
Based on the project's public README and repository metadata.