agentacct is a local‑first tool that reads the log files of AI coding assistants (Claude Code, Codex, OpenCode, Hermes, etc.), joins them with the actual commands, file edits and tool calls the agents performed, and produces a per‑task “Work Receipt” showing actions, token cost and an evidence‑tier (agent claim → self‑check → hook‑observed → CI). All data stays on the machine; a macOS GUI, a terminal dashboard (`agentacct tui`) and a local JSON API let you view or query the receipts. Install via a notarized macOS app or with `pipx` on macOS/Linux, run `agentacct onboard` once, then use the UI to audit and trust your AI‑generated work.

A human-centric framework for building and manage real-life AI and ML systems, streamlining the transition from prototyping to production deployment.

A dynamic workflow runtime for building observable, reusable agent graphs using a code-first TypeScript approach to control AI agent execution.

A hat-based orchestration framework that keeps AI agents in an iterative loop with automated validation gates until a task is fully completed.

An open-source, self-hosted knowledge base and task workbench that integrates rich-text/Markdown notes with AI RAG capabilities for private data management.

An operating system for embodied AI that decouples models and skills from robot hardware, enabling a single AI brain to be deployed across heterogeneous robot bodies.

Vibing Steampunk (vsp) is a Go CLI that bridges SAP’s ADT APIs (via RFC or HTTPS) to let AI assistants and developers read, write, debug, and analyse ABAP code and data—without installing anything on the SAP server. It supports full debugging (breakpoints, stepping, variable inspection/modification), AMDP debugging, dump analysis, job/spool handling, documentation extraction, and SAP‑cluster table decoding, and is testable offline via recorded ADT sessions.

A collection of modular AI workflows (Codex Skills) for Xiaohongshu operations, helping creators systematically optimize their profiles, content planning, and conversion paths.

A GPU-accelerated data curation framework for building repeatable pipelines to load, filter, deduplicate, and transform massive text, image, video, and audio datasets for AI training.

A lightweight, automated runner for RWKV models that provides an OpenAI-compatible API and a user-friendly interface for chat, completion, and music composition.

Twinny is a free VS Code extension that adds AI‑driven code completion, an interactive chat sidebar, workspace‑wide context via embeddings, and optional P2P inference sharing. It works with many LLM providers (OpenAI, Anthropic, Mistral, etc.) and can run locally, offering customizable prompts and features like diff view, commit‑message generation, and offline operation.

A specialized plugin for Claude Code and other AI assistants that provides 26 specialist agents and 51 skills to help developers build bug-free Elixir and Phoenix applications.

LLM Scraper is a TypeScript library that lets you extract structured data from any webpage using large language models. It combines browser automation with schema-based prompting to turn raw web content into organized JSON, supporting multiple LLM providers and streaming output.

A tile-based runtime engine for ultra-low-latency LLM inference that minimizes time per output token for massive models on NVIDIA B200 GPUs.

A multimodal data augmentation library that provides over 100 transforms for audio, image, text, and video to improve AI model robustness against real-world internet data.

An automated video distribution pipeline that handles everything from file monitoring and AI-generated captions to multi-platform uploading using browser automation.

A PyTorch-native library that enables bidirectional checkpoint conversion and high-performance training between Hugging Face and NVIDIA Megatron Core.

fractal is a Python library/CLI that builds a tree of autonomous LLM‑agent loops, each running in its own Git worktree and tmux session. Nodes can spawn child nodes, allowing recursive self‑organization to solve complex tasks. The system tracks runs in SQLite, offers a terminal UI dashboard, and supports multiple LLM back‑ends (Claude, Codex, Grok, OpenCode, Oh My Pi). Safety warnings note that agents run without permission prompts, so they should be used on trusted or sandboxed hosts.

TESSERA is an open geospatial foundation model that converts cloud-corrupted satellite time series into compact 128-dimensional embeddings for global Earth representation and analysis.

Vime is an open‑source RL post‑training framework for large language models. It combines the slime/Megatron distributed training stack with vLLM inference (via vllm‑router) to provide high‑throughput rollouts and flexible data‑generation pipelines. Supported models include Qwen, DeepSeek V3, and LLaMA 3. Users can run agentic RL, coding‑agent RL, or custom RL algorithms by configuring Megatron, vLLM, and router arguments. The repo offers quick‑start docs, several ready‑made examples, and a clear code‑reading path for developers.

An enterprise-grade Java backend and management platform for ESP32 smart hardware, providing real-time AI voice interaction, RAG, and IoT device management.

A storage system for Hugging Face Hub that uses chunk-based deduplication to efficiently upload and download large AI models and datasets.

A macOS CLI tool for on-device speech transcription of files, live system audio, and microphone input using Apple's Speech.framework.

A ComfyUI extension that optimizes inpainting by cropping the masked area, processing it at a target resolution, and stitching it back seamlessly into the original image.

A gateway that enables free access to multiple premium LLMs by automating browser logins and driving official web UIs instead of using paid API tokens.

DeepInverse is an open-source PyTorch-based library designed to solve imaging inverse problems using deep learning through a modular framework of operators and algorithms.

A Python library for user-friendly time series forecasting and anomaly detection, offering a unified API for models ranging from classic statistical methods to deep neural networks.

A CLI and MCP server that connects AI agents and users to Toss Securities, providing full official API coverage plus 58 additional internal web features for trading and analysis.

A self-hosted inference server for Apple Silicon that provides OpenAI and Anthropic compatible APIs for running LLMs, VLMs, and image generation models locally.

Local Studio is an open‑source macOS/desktop app that lets you run, manage, and chat with self‑hosted LLM back‑ends (vLLM, SGLang, llama.cpp, MLX). It combines a Bun + Hono controller (model lifecycle, OpenAI‑compatible proxy, GPU metrics) with a Next.js/Electron UI and an integrated Pi coding‑agent. The UI works locally or can point at a remote controller, and a companion mobile app (KittyLitter) can pair to reuse the same sessions. The repo provides full dev‑setup scripts, production build instructions, and a CI‑driven release pipeline.

A local AI agent framework for iOS that runs LLMs on-device to privately manage calendar, health data, and device tasks via natural language.

An open-source SDK for AI agent safety that blocks prompt injections, redacts sensitive data, and scans repositories for AI-targeted threats.

A tool that converts PDF files into structured Markdown and EPUB formats using intelligent layout detection and AI-powered text extraction.

An AI API gateway that provides smart routing, automatic failover, and cost control for Claude Code, Codex, Gemini, and OpenAI upstreams.

A collection of portable skill files that guide AI coding agents through professional software engineering practices like design, planning, and review.

A consolidated super-repository for ROCm libraries that streamlines the development, CI, and integration of multiple GPU-accelerated math and compute libraries.

A production-ready MCP server that connects AI assistants to Nextcloud, enabling LLMs to perform CRUD operations on notes, calendars, and files via natural language.

A comprehensive sample code library and toolset for the Korea Investment & Securities Open API, featuring LLM-optimized snippets, a visual strategy builder, and a backtesting engine.

A macOS menu bar utility that consolidates system toggles and routine tasks into a single switch hub, featuring AI-powered natural language control and custom scripting.

A free, open-source transcription app that converts speech to text locally on your computer using various AI models for privacy and speed.

A pipeline parallel training script that allows for the training of large-scale image and video diffusion models across multiple GPUs.

A machine learning library that parses raw PDF scientific publications into structured XML/TEI documents for high-scale bibliographic data extraction.

An open-source communication protocol suite that enables AI agents to securely authenticate, discover, and collaborate across different platforms using standardized identities and messaging.

A code-first Java toolkit for building, evaluating, and deploying AI agents with flexible orchestration and tight integration with Google Cloud services.

A command-line productivity tool that uses LLMs to generate shell commands, code snippets, and documentation directly in the terminal.

A 3D interactive visualization tool that displays the network topology and inference process of GPT-style Large Language Models.

A runtime substrate for AI agents that provides secure sandboxing, reversible execution traces, and a review-and-approve workflow for all filesystem changes.

A mobile device automation framework for iOS and Android that brings Playwright's API and developer experience to mobile apps, specifically optimized for both E2E testing and AI agent control.