Fresh Ai Tools articles, tutorials, technical notes, and tool updates from StableLearn.
A beginner-friendly guide to installing OpenClaw on Mac and Windows. One command for Mac, WSL2 or PowerShell for Windows. Covers API setup, Telegram/Feishu integration, and common troubleshooting.
GLM-OCR: 0.9B OCR model tops OmniDocBench V1.5 (94.62). Fast inference at 1.86 pages/sec, excels in formulas, tables & complex docs. Open-source with vLLM/SGLang/Ollama support.
Fun-ASR-Nano-2512 (released 2025-12-15) is a real-time ASR model supporting 31 languages and 7 dialects. This guide covers installation, AutoModel inference, and VAD usage.
Released on 2025-12-15, CosyVoice 3.0 features 9 languages, 18+ dialects, and pronunciation inpainting. This guide covers model setup, inference, and full local deployment.
HunyuanOCR: Tencent 1B OCR model beats DeepSeek-OCR (92% vs 10% on cards). Supports 100+ languages, vLLM OpenAI API deployment. Full code examples included.
BettaFish is an innovative multi-agent public opinion analysis system that supports one-click Docker deployment. This article provides a comprehensive guide on how to quickly deploy BettaFish using Docker, including environment configuration, LLM integration, usage instructions, and troubleshooting.
DeepSeek OCR App: Open-source React+FastAPI text recognition tool with 4 OCR modes - plain text, image description, content search & custom prompts. Docker deployment, GPU acceleration, drag-drop upload. Complete tutorial for developers and enterprises.
PaddleOCR-VL Tutorial: Baidu 0.9B AI model for 109-language OCR. Parse text, tables, formulas & charts with Docker/Python API. SOTA performance beats GPT-4V.
vLLM now supports Xiaohongshu's dots.ocr for free multilingual OCR. Follow our tutorial to deploy this powerful model in two steps for parsing text, tables, and formulas.
GLM-Image is the first open-source industrial-grade discrete autoregressive image generation model with hybrid architecture. Excels in text rendering and knowledge-intensive scenarios with 97.88% Chinese text accuracy.
