Hermes Agent (NousResearch) sandboxed by NVIDIA OpenShell — hardware-enforced network/filesystem/syscall policy, full memory + gateway stack

The best-benchmarked open-source AI memory system. And it's free.

A minimal fine-tuning repo for LFM2, fully built on Open Source.

Local knowledge graph for Claude Code. Builds a persistent map of your codebase so Claude reads only what matters — 6.8× fewer tokens on reviews and up to 49× on daily coding tasks.

The repo is finally unlocked. enjoy the party! The fastest repo in history to surpass 100K stars ⭐. Join Discord: https://discord.gg/5TUQKqFWd Built in Rust using oh-my-codex.

Local models support for Microsoft's graphrag using ollama (llama3, mistral, gemma2 phi3)- LLM & Embedding extraction

Official code base for LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels

Lightweight coding agent that runs in your terminal

Leo optimizer, variation of Muon that runs faster

Build, run, manage agentic software at scale.

AlphaEdit: Null-Space Constrained Knowledge Editing for Language Models, ICLR 2025 (Outstanding Paper)

Dr. Zero Self-Evolving Search Agents without Training Data

Keras implementation of Representation Learning with Contrastive Predictive Coding

mHC kernels implemented in CUDA

Qwen-Image-Layered: Layered Decomposition for Inherent Editablity

DigitalPlat FreeDomain: Free Domain For Everyone

[ECCV 2024] Official implementation of the paper "Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection"

Reference PyTorch implementation and models for DINOv3

Renderer for the harmony response format to be used with gpt-oss

[ICLR 2024] Efficient Streaming Language Models with Attention Sinks

[NeurIPS 2025] The official repository for our paper, "Open Vision Reasoner: Transferring Linguistic Cognitive Behavior for Visual Reasoning".

The code for PixelRefer & VideoRefer

RoboBrain 2.5: Advanced version of RoboBrain. Depth in Sight, Time in Mind. 🎉🎉🎉

Generative Omnimatte (CVPR 2025)

Knowledge Engine for AI Agent Memory in 6 lines of code

Pixel-Level Reasoning Model trained with RL [NeuIPS25]