Projects Recently Starred by Experts

Starred 2 hours ago by

Wei-Lin Chiang(Cofounder of LMArena). Also starred by

Luis Capelo(Cofounder of Lightning AI),

Zhiqiang Xie(Coauthor of SGLang),

Chip Huyen(Author of "AI Engineering", "Designing Machine Learning Systems"), and

1 more.

KernelBench by ScalingIntelligence

0.5%

809

Benchmark for LLMs generating GPU kernels from PyTorch ops

Evaluates LLM-generated CUDA kernels for correctness and performance (speedup vs. PyTorch).
Includes single-kernel ops, fused patterns, and full model architectures (MobileNet, VGG, Mamba).
Uses "fast_p" metric: fraction of correct kernels with >p speedup.
Provides scripts for evaluation, analysis, and baseline timing.

Created 1 year ago

Updated 1 month ago

Starred 13 hours ago by

Johannes Schickling(Cofounder of Prisma). Also starred by

David Cournapeau(Author of scikit-learn) and

Justin Torre(Cofounder of Helicone).

cmux by manaflow-ai

157.4%

950

Parallel coding agent CLI manager

Run multiple AI coding assistants like Claude, Codex, and Gemini concurrently.
Spawns isolated VS Code workspaces for verifiable, fast agent task execution.
Supports cloud or local Docker environments for flexible deployment.
Integrates git diff, terminal, and dev server previews within agent workspaces.

Created 3 weeks ago

Updated 12 hours ago

Starred 15 hours ago by

Jonathan Ragan-Kelley(Professor at MIT). Also starred by

Travis Fischer(Founder of Agentic),

Shane Thomas(Cofounder of Mastra),

Abhi Aiyer(Cofounder of Mastra), and

2 more.

get-shit-done by gsd-build

19.5%

17k

AI development system for reliable code generation

Spec-driven development, meta-prompting, and context engineering for AI code.
Subagent orchestration with 200k token contexts per task ensures consistent AI quality.
Utilizes XML prompt formatting and atomic Git commits for structured, traceable development.
CLI tool automates project setup, roadmap planning, and autonomous task execution.

Created 2 months ago

Updated 2 days ago

Starred 21 hours ago by

Andy Konwinski (Cofounder of Perplexity, Databricks). Also starred by

Logan Kilpatrick(Product Lead on Google AI Studio),

Philipp Moritz(Cofounder of Anyscale),

Shyamal Anadkat(Research Scientist at OpenAI), and

3 more.

harbor by laude-institute

9.3%

696

Harness for AI agent evaluation and RL optimization

Evaluate diverse agents (Claude Code, Codex) and LLMs against benchmarks.
Create, share, and run parallel experiments across thousands of environments.
Generate agent rollouts for Reinforcement Learning optimization tasks.
Supports cloud execution via Daytona and Modal providers.

Created 6 months ago

Updated 19 hours ago

Starred 22 hours ago by

Jason Knight(Director AI Compilers at NVIDIA; Cofounder of OctoML). Also starred by

Nir Gazit(Cofounder of Traceloop),

Travis Fischer(Founder of Agentic),

Matt McClure(Cofounder of Mux), and

6 more.

nanoclaw by qwibitai

30.5%

10k

Personal AI assistant for secure, containerized operation

AI agents run in isolated containers (Docker/Apple Container) for robust security.
AI-native interface guides setup, customization, and debugging via Claude Code.
Extensible via code skills, supporting scheduled tasks and multi-channel I/O (e.g., WhatsApp).
Lightweight Node.js app designed for deep understanding and user-driven code customization.

Created 3 weeks ago

Updated 21 hours ago

Starred 1 day ago by

Lysandre Debut(Chief Open-Source Officer at Hugging Face). Also starred by

Tim J. Baek(Founder of Open WebUI),

Julien Chaumond(Cofounder of Hugging Face), and

Georgi Gerganov(Author of llama.cpp, whisper.cpp).

LlamaBarn by ggml-org

3.9%

960

Easy local LLM deployment on Mac

Run diverse LLMs with one click using a native macOS menu bar app.
Features a chat UI and REST API, built upon the llama.cpp inference engine.
Automatic hardware configuration ensures optimal performance and stability for local models.
Tiny (~12MB) app with curated catalog, simplifying LLM access without system-wide installs.

Created 7 months ago

Updated 4 days ago

Starred 1 day ago by

Wei-Lin Chiang(Cofounder of LMArena). Also starred by

Wing Lian(Founder of Axolotl AI) and

Yaowei Zheng(Author of LLaMA-Factory).

GamingAgent by lmgame-org

0.3%

866

SDK for LLM/VLM gaming agents, enabling model evaluation via games

Supports agents for platformer/Atari games like Super Mario Bros, Sokoban, 2048, Tetris.
Uses OpenAI, Anthropic, Gemini, Deepseek models; vision-text modalities.
Includes specialized workers for evidence, memory, dialog in Ace Attorney.
Allows custom policies, concurrency strategies, and game-specific tuning.

Created 1 year ago

Updated 3 months ago

Starred 1 day ago by

Dan Guido(Cofounder of Trail of Bits).

DeepAudit by lintsinghua

0.9%

LLM-powered code auditing for enhanced quality and security

Analyzes code for security, performance, and quality issues across 10+ languages.
Integrates with GitHub/GitLab for automated CI/CD code reviews.
Supports 10+ AI models including local Ollama, with clear "What-Why-How" explanations.
Built with React/TypeScript, offering local or cloud data storage options.

Created 5 months ago

Updated 1 week ago

Starred 2 days ago by

Kevin Hou(Head of Product Engineering at Windsurf). Also starred by

Tim Suchanek(Founder of expand.ai),

Gregor Zunic(Cofounder of Browser Use),

Abhi Aiyer(Cofounder of Mastra), and

1 more.

superset by superset-sh

12.5%

Parallel CLI coding agent management terminal

Run dozens of CLI agents concurrently, managing tasks and switching contexts efficiently.
Leverages git worktrees and a unified terminal system for isolated, parallel agent execution.
Integrated diff view with editor for inspecting and editing agent outputs directly within the tool.
Automates setup and monitors agent progress, notifying users when tasks are ready.

Created 4 months ago

Updated 13 hours ago

Starred 2 days ago by

Lysandre Debut(Chief Open-Source Officer at Hugging Face). Also starred by

Travis Fischer(Founder of Agentic) and

Luis Capelo(Cofounder of Lightning AI).

fast-agent by evalstate

0.4%

SDK for agent and workflow creation with MCP feature support

Declarative syntax for composing prompts and MCP servers.
Supports Anthropic and OpenAI models, including multimodal inputs (images, PDFs).
Includes passthrough/playback LLMs for rapid development and testing.
Features agent chaining, parallel execution, evaluator-optimizer workflows, and routing.

Created 1 year ago

Updated 1 day ago

...