Welcome to the LocalAI organization — home of open-source, privacy-first AI inference tools that run locally on your own hardware.
We build lightweight C++/GGML runtimes, neural audio/video pipelines, and model tooling so you can run state-of-the-art AI without sending data to the cloud.
| Repository | Description | Language | Stars |
|---|---|---|---|
| privacy-filter.cpp | OpenAI's privacy filter NER model architecture implemented in a minimal C++/GGML runtime | C++ |  |
| vibevoice.cpp | C++ port of Microsoft VibeVoice built on ggml | C++ |  |
| rf-detr.cpp | RT-DETR object detection implemented in C++ with ggml | C++ |  |
| ced.cpp | GGML port of CED — SOTA-tier audio-tagging model that classifies everyday sounds | C++ |  |
| free-splatter.cpp | Turn images into 3D scenes — GGML port of FreeSplatter for sparse-view 3D reconstruction | C++ |  |
| Repository | Description | Language | Stars |
|---|---|---|---|
| apex-quant | Adaptive Precision for EXpert Models: MoE-aware mixed-precision quantization | Shell |  |
| LocalVQE | Lean neural real-time acoustic echo cancellation with soft delay estimation — GGML and PyTorch inference | C++ |  |
| localai-realtime-demo | LocalAI realtime demo | Go |  |
| italian-asr | Open-source models that turn Italian audio into text, locally, with optional speaker diarization | N/A |  |
💡 All our C++ projects are built on GGML and designed to run inference locally without cloud dependencies.