Neural Research Lab
An independent applied research lab building tools for edge inference, private assistants, labeling, and training. Offline-first, real licenses, and designed for the hardware you already own.
- 7
- 01 · Projects
- 5
- 02 · open-source
- 4k+
- 03 · github stars
Projects that ship.
Every line below is a project you can clone, run, or deploy today. Sorted roughly by how often we see it used in the wild.
01 →Nômopen-sourcelanguage modelOpen-source Python toolkit for Vietnamese AI: RAG over docs, diacritic restoration, segmentation, OCR, document parsing. Local-first, multi-backend (Ollama, llama.cpp, OpenAI, Anthropic). Apache 2.0.
PythonVN AIApache 2.0since 2026
02 ↗AnyLabelingopen-sourcevisionEffortless AI-powered data labeling with Segment Anything and YOLO. Auto-annotation, offline capable, exports to common formats.
PythonPyQtYOLOSAM★3.3ksince 2023
03 ↗Llama Assistantopen-sourceassistantsPrivacy-first desktop AI assistant. Voice, multimodal, RAG, screen OCR, "Hey Llama" wake word — fully offline on your device.
PythonLlama 3.2DeepSeek R1RAG★532since 2024
04 ↗EdgeVoxpreviewassistantsOn-device voice-agent framework. STT → LLM → TTS pipeline that runs entirely offline, with stop-word halt for safety. ROS2-native; targets robots, desktop apps, and embedded systems.
PythonROS2STTLLMTTS★5since 2026
05 ↗AnyLearningopen-sourceplatformOpen-source, offline-first no-code AI platform for data labeling, computer vision, tabular ML, and local model training.
PythonPyTorchNext.jsApache 2.0since 2026
06 ↗DaisyKitopen-sourcetoolkitPlug-and-play AI toolkit for edge devices. Face detection, pose estimation, background matting — no deep learning knowledge required.
C++PythonONNXTFLite★110since 2022
07 ↗LilyLingoliveedtechAI-powered language learning with virtual friends. Vocabulary, pronunciation practice, and interactive role-play — anywhere, anytime, any device.
ReactSpeech AINLPsince 2025