Starred repositories
SLAM Dunk: A recipe for long horizon egocentric SLAM in the wild
[NeurIPS 2025] Tracking and Understanding Object Transformations
Goekdeniz-Guelmez / caramel
Forked from naklecha/caramelcaramel-1, a full-duplex speech-to-speech model trained only on audio in MLX
Duo: A Reciprocal World Model for Full-Duplex Interaction
A library of behaviors for agents to build the next generation of user interfaces.
Write an app once. It runs natively on the web, macOS, iOS, and Linux, and an AI agent can build it, run it, see it, and test it on every one of them.
Code for Latent Sparse Attention: Towards Subquadratic Ultra-Long Context Attention
WeVisOne-VL is a powerful VLM by the WeChat Vision Team at Tencent, unifying object-level and pixel-level perception from points and boxes to masks and alpha mattes.
A framework for building robust RL environments to train aligned AI.
Towards a More Autonomous Agent System
Every past session, subagent, and workflow -- queryable by your agent, browsable by you
A framework for building decision-model harnesses for any kind of task. Your program offers the model a short list of next steps, and the model picks one at a time until it can answer.
BELL: the open source personal assistant powered by Browserbase, Stripe, Vercel, and Linq.
Professional Standalone Music Production Center with 7" Touch Screen, Sampler, Touch Strip and Music Tribe JAM DAW
YC F26. Turn any website into an API / MCP in minutes. Open source core integration minting harness. MIT.
Steering-first neuron explorer for Qwen3.5-4B: push every MLP neuron, keep every rollout, browse the results
A context-compression plugin for month-long single sessions (billions of tokens) and small context windows (a 100K context is enough), billion-context is all you need .
A lightweight sandboxing tool for enforcing filesystem and network restrictions on arbitrary processes at the OS level, without requiring a container.
DeepInfer: High-performance, large-scale inference framework for DeepSeek models
Moshi is a speech-text foundation model and full-duplex spoken dialogue framework. It uses Mimi, a state-of-the-art streaming neural audio codec.
Turn your regular JavaScript code into a neural net.
Reverse engineer anything with agents, from app behavior down to native binaries.
A lightweight, native Swift 6 bridge integrating TypeSafe AI's Jev System One decision model into Apple's Foundation Models framework.
LiteReality-Agent: Turn the real world into simulation-ready environments.
12,000+ open source STEP parts for your next CAD project
Argus is an open source dataset-and-format-agnostic harness for the densest annotation of robot learning data.