edge-ai

Fully local voice AI for iOS

On-device AI SDK for Flutter — LLM inference, vision, STT, TTS, image generation, embeddings, RAG, and function calling. Metal GPU on iOS/macOS.

Gemma models running 100% local on an 8 GB Jetson Orin Nano: voice assistants with RAG, multi-agent NPCs, live EN-JA translation, and a Gemma 4 voice + vision agent that decides when to look through the webcam. The demos from the Gemma 2 launch at Google Tokyo.

On-device VAD / streaming STT / TTS / diarization in C++17 (ONNX + LiteRT) with a voice-agent pipeline. Linux, Windows, Android.

Android native AI inference library, bringing text, image, video, STT, TTS inference

Distributed 56M-parameter LLM inference across 3 ESP32-S3 boards via ESP-NOW , Split-PLE + KV cache, fully offline.

POC visual search with smart glasses and Qdrant Edge.

Official Implementation of VisualClaw: A Real-Time, Personalized Agent for the Physical World

From-scratch voice agents in Python: end-to-end speech pipelines, runnable chapters, and a small shared library. Local models, explicit streaming behavior.

A 30-day build challenge for people who can execute under pressure. You will build one real project every day across software, AI, systems, and applied engineering. Miss one day, and you are out. This is not a course. This is not passive learning.

🔥 Alternative to Ollama — multi-model serving with sub-ms model switching · CPU-only 20B inference for Edge AI · llama.cpp + stablediffusion.cpp

Light-MER for efficient multimodal emotion recognition with sub-1B MLLMs.

Export and Run Hugging Face Transformers Models on Rockchip NPUs

LLM-free, layout-aware PDF chunking for RAG pipelines. Preserves document structure via regex and font heuristics. Written in pure Rust.

High-performance, Hardware-First Vector Database engine built in Zig and Go. Sub-millisecond HNSW search with 64-bit metadata filtering and 1.8MB binary footprint.

🧠 Predictable, embedded multimodal vector database & offline RAG engine for Edge AI (ARM64 / Zero-OOM). Built in C++20.

Local AI load balancer for Ollama and MLX fleets. Auto-discovery, smart routing, OpenAI-compatible API, zero config. Perfect for Mac Minis & Studios.

AI-Inference-Managed-by-AI: Go binary for managing AI inference on edge devices

Persistent semantic memory for AI agents on Raspberry Pi 5 — local Qdrant + MCP, no cloud, ~3s per query

Running small but capable language models entirely offline