Software · SDK qb
Your model
TensorFlow
TFLite
ONNX
KERAS
PyTorch
qb CompilerManual
Translates deep learning models into Mobilint's executable format, MXQ.
Open manual →
Access pre-compiled models →
.mxq
qb RuntimeManual
Convenient and fast inference using internal optimized logic. C++ and Python interfaces.
Open manual →
Mobilint NPU
REGULUS devices ship with the essential SDK stack — driver and runtime library — pre-installed. Manual driver/runtime installation applies to ARIES-based devices only.
Hardware
Prerequisites, installation, and platform guides
ARIESManual
One manual for every ARIES-based device — installation, driver, firmware, and platform guides.
MLA100 PCIe Card
MLA100 MXM
MLX-A1 Edge AI PC
Open manual →
REGULUSComing soon
Ultra-low-power SoC device. Devices ship with the SDK stack pre-installed — documentation will be available in a later phase.
Resources
Models, framework integrations, downloads, and community
Mobilint on GitHub — open-source integrations, models, tools, and tutorials for running your stack on the Mobilint NPU.
github.com/mobilint ↗
mblt-model-zoo ↗490+ models
Pre-compiled open-source models for vision, language, and multimodal AI, ready to run as
.mxq.
vllm-mblt ↗
Serve LLMs on the NPU through the vLLM plugin.
mblt-sdk-tutorial ↗
Hands-on tutorials for SDK qb, from compile to inference.
…and more — trackers, device plugins, and demo apps. Explore the org page to see everything public.