A node.js embedding tool with optional GPU acceleration

Compiles Typescript code into Build Engine CON

Clawmark — durable facts and semantic recall for OpenClaw, derived from the message history you already keep. No duplicate message store; crash-safe by design.

Fastest SIMD-Accelerated Vector Similarity Functions for x86 and Arm

CLI for discovering, inspecting, and running ComfyUI workflows with tag-based overrides

Find which LLMs can run on your system via OpenRouter

Renderer-agnostic core SDK for Omote AI Characters

Intelligent CLI tool with AI-powered model selection that analyzes your hardware and recommends optimal LLM models for your system

Run Hugging Face onnx-community models locally inside pi: registers a chat provider for ONNX text-generation models and a set of tools (embeddings, classification, ASR) backed by @huggingface/transformers and onnxruntime-node.

// Half-precision floating-point format // bfloat16

On-device 14-class keyword spotter in the browser — WebAssembly runtime + streaming Conformer, ~1.5 MB total bundle, no server required

High-quality text-to-speech for the web, support Chinese, web播放文本到语音，支持中文语音.

Run LLMs locally in your terminal. Supports custom .pt GPT-2 and LLaMA decoder checkpoints (auto-handles compact key naming, tied embeddings, and split-halves RoPE) plus GGUF models, with a live dashboard of throughput, VRAM, and chat.

Remove background from images using AI

React Native binding of whisper.cpp

Web browser version of StarCoder.cpp

Node.js translation library powered by Transformers.js with ONNX Runtime

Runtime-agnostic translation core powered by Transformers.js

CLI preflight checks for model VRAM, CUDA, GPU compatibility, and quantization risk.

Flower Intelligence: Open-Source On-Device AI with optional Confidential Remote Compute.