GH browser-use 12h ago 95

browser-use/browser-use

Browser automation framework for AI agents. Enables LLMs to interact with web interfaces and automate online tasks programmatically.

AI agentsDeveloper tools
GH OpenHands 12h ago 95

OpenHands/OpenHands

AI-driven development platform enabling agents to write and execute code. Turns LLMs into autonomous developers using CLI and orchestration.

AI agentsDeveloper tools
GH langgenius 12h ago 96

langgenius/dify

Production platform for agentic workflow development with low-code/no-code interface. Enables building and deploying agent-based automation at scale.

AI agentsworkflow platform
DT Vansh Uttam 13h ago 86

🎯MCP vs Direct API Calls

Model Context Protocol (MCP) vs direct APIs for LLM integration, designed for developer abstraction.

LLMdeveloper tools
DT Vansh Uttam 13h ago 84

How caching helps in LLM Application?

Explains caching mechanisms for LLM applications to reduce latency and costs. Performance optimization technique for production LLM deployments.

LLM applicationsoptimization
GH ultralytics 13h ago 82

ultralytics/ultralytics

Ultralytics YOLO object detection framework with PyTorch backend. Production-ready ML models for vision tasks with deployment options.

ML researchcomputer vision
GH huggingface 14h ago 90

huggingface/transformers

Model framework for inference and training across text, vision, audio, and multimodal. Foundational library for LLM applications and agent development.

LLM applicationsOpen source
HN manthangupta109 14h ago 80

Views on Mastra's SOTA Memory

Technical analysis of Mastra's memory implementation for AI agents, evaluating SOTA claims and production robustness. Directly relevant to LLM applications.

LLM agentsmemory
HN buttersmoothAI 15h ago 79

Zero State Architecture deep dive

Architecture eliminating context drift in long-horizon LLM tasks through stateless design preventing coherence loss and recursive loops.

AI agentsmachine learning research
HN broyojo 16h ago 78

Show HN: Running an LLM Inside Scratch

LLM inference implementation running quantized Llama2 model inside Scratch/TurboWarp by compiling C code to Scratch blocks.

LLM applicationsdeveloper tools
GH gradio-app 17h ago 75

gradio-app/gradio

Gradio framework for building and deploying ML applications via Python. Simplifies UI creation for machine learning models and services.

developer toolsLLM
DT Ned C 17h ago 70

How to Write .cursorrules That Actually Work

Framework for writing .cursorrules configuration files that Cursor AI editor actually follows. Practical guide for AI-assisted development.

developer toolsLLM
GH hiyouga 18h ago 90

hiyouga/LlamaFactory

Unified framework for fine-tuning 100+ LLMs and VLMs. Supports LoRA, QLoRA, instruction tuning, and RLHF for efficient model adaptation.

LLM applicationsOpen source
GH infiniflow 19h ago 95

infiniflow/ragflow

Open-source RAG engine with agentic workflow capabilities. Combines retrieval-augmented generation with agent orchestration for enhanced LLM context.

AI agentsLLM applications
GH scikit-learn 19h ago 55

scikit-learn/scikit-learn

Classical machine learning library in Python. Foundational ML tool but predates modern LLM applications and agent frameworks.

ML researchDeveloper tools
DT Dev In Progress 19h ago 81

How Transformer Architecture Powers LLMs

Technical deep-dive into how transformer architecture functions in LLMs, beyond surface-level explanations.

ML researcheducation
LB alexzhang13.github.io via adsouza 20h ago 75

Recursive Language Models

Explores recursive approaches in language models, relevant to understanding advanced LLM architectures and how models can process hierarchical information iteratively.

LLM researchmachine learning
GH hacksider 21h ago 72

hacksider/Deep-Live-Cam

Open source Python project for real-time face swapping and deepfake generation using GANs. Practical implementation of deep learning techniques.

open source AIcomputer vision
GH unslothai 21h ago 95

unslothai/unsloth

Fine-tuning and reinforcement learning library optimized for LLMs. Reduces training time and memory for Llama, Qwen, Gemma, DeepSeek and other models.

LLM applicationsOpen source