Keep Calm and Read the Friendly Manual :-)
Description:
PicoClaw is an ultra-lightweight, Go-powered personal AI assistant designed to run on low-end hardware (under 10MB RAM). It supports multiple LLM providers, runs on Linux/macOS/Windows, and deploys across Telegram, Discord, WhatsApp, LINE, and more. It emphasizes local execution and security, with a small footprint and modular architecture.
Description:
A local, self-hosted web app for storing, organizing, and refining AI prompts for image, music, and video generation. It supports prompt variants, grouping, full-text search, tagging, and optional local LLM-assisted drafting, with all data stored locally.
Description:
SmartClipboard is a local, privacy-first clipboard workspace that rewrites, summarizes, reformats, and extracts structured information using local Ollama-based LLMs. It stores clipboard history locally (SQLite), supports a Composer workspace, AI actions, and JSON-driven personas, with no cloud APIs and data staying on the user’s machine.
Description:
Trigger.dev is an open-source platform for building AI workflows in TypeScript. It enables long-running tasks with retries, queues, observability, and elastic scaling, and lets you build and deploy AI agents and durable workflows.
Description:
IronClaw is an OpenClaw‑inspired, open‑source AI assistant written in Rust. It emphasizes privacy and security with local encrypted data storage, a WASM sandbox for tools, and defense‑in‑depth measures. It supports multi‑channel access, plugin expansion, and self‑expanding capabilities.
Description:
Fast, small, and autonomous AI agent infrastructure. ZeroClaw is a Rust-based private control plane with swappable providers, channels, tools, and memory, designed for lean runtimes (<5MB RAM) across ARM, x86, and RISC-V.
Description:
LTX-2 is an open, production-grade DiT-based model for synchronized audio–video generation, delivering high-fidelity outputs with multiple performance modes, API access, and open design for production workflows.
Description:
AnyCrawl is a high-performance Node.js/TypeScript crawler that turns websites into structured data for LLMs, enabling SERP scraping, page extraction, and scalable site crawling.
Description:
Agent Lightning is an open-source framework to train AI agents with reinforcement learning. It enables near-zero-code integration with major agent frameworks, supports multiple algorithms, and uses a central LightningStore to coordinate tasks, prompts, and rewards across runners and environments.
Description:
whisper.cpp is a high-performance C/C++ port of OpenAI’s Whisper for offline transcription. It runs on Apple Silicon with NEON/Metal and on x86 with AVX, supports quantization, Vulkan and multiple backends, and includes a lightweight ggml-based engine for cross-platform use (macOS, iOS, Linux, Windows, WASM).