#edge-ai

Agents delivered by light: an LLM compiles a tiny program off-chip and beams it by screen-light to a radio-less Arduino that runs it forever, offline. Live at beam-claw.com

Latency-Optimized Heterogeneous AI Accelerator for Vision Transformers (ViT)

OpenAI-compatible HTTP API for Needle 2, Cactus Compute's 45M-parameter tool-calling model. Drop-in for any OpenAI client.

Long-horizon reliability benchmark for industrial edge agents - do self-improving agents get better or corrupt over 1000+ episodes on constrained hardware?

Model-agnostic substrate for running small local LLMs well — code agents, chat, image/video, speech. Small capable main model + tiny fast worker.

Quantization, sparsity, and behavioral robustness of Qwen LLMs under edge hardware constraints

TUI academia: math → Edge AI. Neovim keys. Rank Edge Mage se merece.

Local inference server in pure Mojo. Speaks OpenAI, Anthropic, and MCP. Models are OCI artifacts. Same source runs on CPU, NVIDIA, AMD, and Apple GPUs. Apache-2.0 all the way down.

Local object detection hub utilizing OpenVINO optimization for real-time video analytics, running an asynchronous multi-threaded pipeline with a custom web UI.

Source-available view of Cubecloud OS — a sovereign-intelligence, local-first AI operating layer for private work on your own machines.

Provider-neutral access & supply layer for open AI: run open models locally (SIN) and route paid, private, verifiable inference across decentralized providers (SIP-AI). DecentralizeAI hackathon entry.

Nimbus is a lightweight, pull-based edge AI workload orchestrator.

Anivar NVR — a free, open-source Blue Iris alternative with on-device AI: faces, plates, sounds and plain-language search. A desktop app, no Docker.

🥖 Offline sourdough bread baking assistant powered by an on-device micro-LLM on ESP32-S3 with OpenAI-compatible WiFi API