Repository History

MarkLLM: An Open-Source Toolkit for LLM Watermarking

MarkLLM is an open-source toolkit designed to simplify the research and application of watermarking technologies for large language models (LLMs). It offers a unified framework for implementing various watermarking algorithms, alongside robust visualization and comprehensive evaluation tools. This toolkit helps researchers and the broader community understand and assess the authenticity and origin of machine-generated text.

EasyWhisperUI: A Cross-Platform Desktop App for Whisper Model Transcription

EasyWhisperUI is a fast, local desktop application designed for transcribing audio and video using the Whisper model. It offers GPU acceleration across Windows, macOS, and Linux, providing a user-friendly interface for various transcription tasks. The application supports features like live transcription, batch processing, and translation, making it a versatile tool for media processing.

OrbitDB: Peer-to-Peer Databases for the Decentralized Web

OrbitDB is a serverless, distributed, peer-to-peer database designed for the decentralized web. It leverages IPFS for data storage and Libp2p Pubsub for automatic synchronization, ensuring eventual consistency through Merkle-CRDTs. This makes OrbitDB an excellent choice for p2p, decentralized, blockchain, and local-first web applications, offering various database types like event logs, documents, and key-value stores.

Dexter: An Autonomous Agent for Deep Financial Research

Dexter is an autonomous financial research agent designed to think, plan, and learn while performing analysis. It leverages task planning, self-reflection, and real-time market data to tackle complex financial questions. This project provides a powerful tool for in-depth financial exploration, emphasizing its educational and informational purposes.

PixelRAG: Pixel-Native Search for Visual Retrieval-Augmented Generation

PixelRAG revolutionizes search by enabling pixel-native retrieval, moving beyond traditional text parsing. It renders documents as screenshots, preserving visual context like tables and charts, which is crucial for accurate answers from reader models. This allows for searching any document based on its visual appearance, not just its textual content.

Recordly: Create Polished Demo Videos with Ease on Mac, Windows, and Linux

Recordly is an open-source screen recorder and editor designed for creating professional demo videos, walkthroughs, and product videos without requiring advanced editing skills. It offers powerful features like auto-zooms, cursor polish, webcam overlays, and a comprehensive timeline editor, making it accessible across macOS, Windows, and Linux. This tool streamlines the video creation process, allowing users to produce high-quality content efficiently.

stablyai-orca

Loop Library: Practical Repeatable AI-Agent Workflows

The Loop Library is a GitHub repository offering reusable AI agent workflows for various domains like engineering, content, and design. It introduces the concept of "loops," which are structured, repeatable instructions that guide AI agents through multi-step tasks. This skill enables agents to learn from results, adapt, and complete complex tasks more reliably than with one-shot prompts.

Agent-Reach: Empower Your AI Agents with Internet Access, Zero API Fees

Agent-Reach is a powerful GitHub repository that equips AI agents with the ability to access and search the entire internet, including platforms like Twitter, Reddit, YouTube, and Bilibili. It provides a streamlined CLI experience, eliminating the need for complex API configurations and associated fees. This project ensures your AI agent can "see" and interact with web content effortlessly.

OpenPencil: The AI-Native, Open-Source Figma Alternative Design Editor

OpenPencil is an innovative AI-native design editor, serving as a powerful open-source alternative to Figma. It supports .fig files, integrates AI for design creation, and provides a fully programmable toolkit with a headless Vue SDK. This project emphasizes real-time collaboration and local data control, making it a compelling choice for designers and developers seeking flexibility and ownership.

PinchTab: High-Performance Browser Automation for AI Agents

PinchTab is a high-performance browser automation bridge and multi-instance orchestrator, designed to give AI agents direct control over Chrome. Built in Go, it offers advanced stealth injection, real-time dashboards, and token-efficient web interaction. It supports both headless and headed modes, enabling robust and secure automation workflows for various applications.

Pic Smaller: Intelligent Online Image Compression for Web and Mobile

Pic Smaller is an intuitive and powerful online tool designed for intelligently compressing various image formats, including JPEG, PNG, WEBP, AVIF, SVG, and GIF. It offers a user-friendly experience with automatic compression and customizable output options, all while ensuring complete safety through purely local processing without server-side logic.