Back to Tool
G

GenieX

Tool Oct 17, 2026

Qualcomm's developer-preview on-device inference runtime for Snapdragon, exposing GGUF and chipset-specific model backends through CLI, Python, Android and a local API server.

Visit primary link 2 sources attached
Qualcomm's GenieX architecture diagram showing client interfaces, a shared SDK, separate llama.cpp and QAIRT plugins, and CPU, GPU and NPU backends.
Author-published architecture diagram in the v0.9.1 README. It illustrates the existing backend layout, not a benchmark, an AITrending test or evidence that every model works on every backend. Full diagram preserved. Source: Qualcomm Technologies / GenieX contributors
self-hostedagent-ecosystemai-agents

What happened lately

GenieX timeline

GenieX is a standalone local inference runtime for Qualcomm Snapdragon devices, rather than a model provider or a coding-agent harness. Its C SDK is exposed through CLI, Python, Kotlin/Java, Docker and an OpenAI-compatible server. The llama.cpp backend consumes GGUF models and supports CPU, Adreno GPU and Hexagon NPU execution; the separate QAIRT backend consumes chipset-specific compiled bundles for the NPU.

The tagged README labels GenieX a developer preview and lists Windows ARM64, Android and Linux ARM64 platforms. The project code has a BSD-3-Clause license, with additional Qualcomm terms and separately applicable model licenses. Version 0.9.1 packages a Microsoft-signed HTP catalog and FastRPC loader and library-path fixes.