Categories
- All Posts 549
- Practical Open Source Projects 478
- Tutorial Articles 22
- Online Utilities 13
- AI news 7
- Tiny Startups Showcase 7
- Claude Code Skills 6
- Prompt Templates 5
- Hugging Face Spaces 3
- OpenClaw Use Cases 3
- LLM Learning Resources 1
- Online AI Image Tools 1
- OpenClaw Master Skills Collection 1
- Rust Training Resources 1
- AI Short Drama Tools 1
- My Favorites 0
Practical Open Source Projects
Practical Open Source Projects
YouTube Shorts Pipeline v2: AI Automates Everything
Transform a single topic into a published YouTube Short in minutes with this fully automated pipeline. From DuckDuckGo research and Claude script generation to Gemini AI visuals, ElevenLabs voiceover, Whisper captions, royalty-free music, and direct YouTube upload – everything is handled. Discover trending topics from Reddit, Google Trends, and TikTok. Resume capability, retry logic, and just $0.11 per video make it production-ready. 78 tests included.
Auto YouTube Shorts Pipeline: Topic to Published Video
Transform any topic into a published YouTube Short in minutes with this fully automated pipeline. From trending topic discovery (Reddit, Google Trends, TikTok) to AI-generated scripts (Claude), visuals (Gemini Imagen), voiceovers (ElevenLabs), burned-in captions, music ducking, and direct YouTube upload. v2.1.0 features resume capability, retry logic, 78 tests, and costs just $0.11 per video. Perfect for creators scaling short-form content.
MiniMax Skills: AI Coding Agent Toolkit
Discover MiniMax Skills – a powerful open-source repository with 15+ specialized skills for AI coding agents. From frontend React/Next.js development to iOS, Android, shaders, and document generation (PDF/DOCX/PPTX/XLSX), this toolkit supercharges Claude, Cursor, Codex, and OpenCode. Generate production-ready code, stunning visuals, and professional documents with MiniMax APIs. Easy installation, active community contributions, and MIT license make it essential for developers.
TurboQuant+: 6.4x KV Cache Compression for LLMs
TurboQuant+ implements ICLR 2026's breakthrough KV cache compression, achieving 4.6-6.4x compression with near q8_0 quality and speed. Features turbo2/turbo3/turbo4 formats, attention-gated Sparse V decoding (+22.8% decode speed), and full llama.cpp Metal integration. Run Qwen 3.5 35B-A3B on M5 Max with 93.9% NIAH retrieval and 1.02x q8_0 prefill speed. Complete Python prototype with 511+ tests and community validation across Apple Silicon, NVIDIA, and AMD.
InkOS: AI Agents That Write Complete Novels
Discover InkOS, the revolutionary CLI tool that automates novel writing with a team of specialized AI agents. From fantasy epics to fan fiction, InkOS handles drafting, 33-dimension auditing, style imitation, and continuity tracking across 7 truth files. Generate 14K+ word chapters with 100% audit pass rates. Supports English novels, multi-model routing, daemon mode, and EPUB export. Perfect for writers seeking AI-powered creativity without losing control.
Chrome-CDP-Skill: AI Access to Your Live Chrome Tabs
Discover chrome-cdp-skill, the game-changing open-source tool that lets AI agents see and control your existing Chrome tabs—no new browser instances needed. Access logged-in accounts, interact with live pages, and automate workflows seamlessly. Works with Chrome, Brave, Edge on all platforms. Install as a pi skill or Node.js module, enable remote debugging once, and unlock commands like screenshots, clicks, typing, and JS evaluation across 100+ tabs. Say goodbye to re-logins and Puppeteer timeouts.
SwiftUI Pro: AI Agent Skill for Smarter Code
Discover SwiftUI Pro, the agent skill that supercharges AI coding assistants like Claude Code, Codex, and Gemini. Created by Paul Hudson of Hacking with Swift, this open-source tool delivers expert guidance on SwiftUI best practices, performance optimization, accessibility, and avoiding common LLM mistakes. Install with a single npx command and instantly improve your SwiftUI code quality. Perfect for iOS developers targeting iOS 26+ and Swift 6.2+ who want production-ready SwiftUI from AI tools.
AI App Store Screenshot Generator for iOS
Generate production-ready App Store screenshots with AI! This open-source skill works with Claude Code, Cursor, and 40+ AI agents. Just describe your iOS app's brand, features, and style—get advertisement-style screenshots exported at all Apple-required resolutions (6.9", 6.5", 6.1"). Supports multi-language, RTL layouts, theme presets, and bulk exports. Install with one npx command and start creating App Store marketing assets instantly.
Helios: 14B Real-Time Video Gen at 19.5 FPS
Discover Helios, the breakthrough 14B parameter video generation model from PKU-YuanGroup that generates minute-scale, high-quality videos at 19.5 FPS on a single H100 GPU. No anti-drifting tricks, no acceleration hacks - just pure architectural innovation. Supports T2V, I2V, V2V, and interactive generation with Day-0 support for Diffusers, SGLang, vLLM-Omni, and Ascend NPU. Run it locally with ~6GB VRAM using group offloading. Complete training code and three model variants (Base, Mid, Distilled) available now.
memory-lancedb-pro: OpenClaw AI Memory Plugin
Transform your OpenClaw AI agents with memory-lancedb-pro – a LanceDB-backed memory plugin that gives agents true long-term memory. Auto-captures conversations, smart-extracts key facts/preferences, and auto-recalls relevant context across sessions. Features hybrid vector+BM25 retrieval, cross-encoder reranking, Weibull decay, and multi-scope isolation. One-click install script makes setup effortless for OpenClaw 2026.3+.