Projects Recently Starred by Experts

Starred 3 days ago by

Jeff Hammerbacher(Cofounder of Cloudera). Also starred by

Lianmin Zheng(Coauthor of SGLang, vLLM),

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

Casper Hansen(Author of AutoAWQ).

MiniCPM by OpenBMB

0.4%

11k

Ultra-efficient LLMs for end devices, achieving 5x+ speedup

Surpasses Phi-3.5, matches Llama3/Qwen2 on benchmarks; supports tool/code interpreter.
Achieves SOTA on Berkeley Function Calling Leaderboard among models <9B parameters.
LLMxMapReduce enables near-infinite context length, outperforming GPT-4 on InfiniteBench.
Supports HuggingFace, vLLM, llama.cpp, SGLang inference frameworks.

Created 2 years ago

Updated 1 week ago

Starred 1 week ago by

Philipp Schmid(DevRel at Google DeepMind). Also starred by

Chris Van Pelt(Cofounder of Weights & Biases).

mcp-go by mark3labs

0.1%

Go SDK for LLM-app integration with external data/tools via Model Context Protocol

Exposes data to LLMs through static/dynamic Resources (files, APIs, DB queries).
Enables LLMs to take actions via Tools (computation, side effects, API calls).
Supports reusable Prompts as interaction templates with system instructions.
Offers session management, request hooks, and tool handler middleware.

Created 1 year ago

Updated 5 days ago

Starred 1 week ago by

Elvis Saravia(Founder of DAIR.AI). Also starred by

Amanpreet Singh(Cofounder of Contextual AI),

Dan Guido(Cofounder of Trail of Bits),

Gregor Zunic(Cofounder of Browser Use), and

1 more.

omnara by omnara-ai

0.2%

Control your AI agents from anywhere

Monitor AI agent progress in real-time across multiple tools like Claude and GitHub Copilot.
Interact with agents via mobile or web, responding to questions and guiding their tasks.
Receive smart notifications only when agent input is required, reducing interruptions.
Built with FastAPI, React, React Native, and PostgreSQL for robust performance.

Created 1 year ago

Updated 23 hours ago

Starred 1 week ago by

Matt Holt(Author of Caddy). Also starred by

Didier Lopes(Founder of OpenBB),

Pietro Schirano(Founder of MagicPath),

Omar Sanseviero(DevRel at Google DeepMind), and

22 more.

web-llm by mlc-ai

0.0%

19k

In-browser LLM inference engine using WebGPU for hardware acceleration

Runs LLMs directly in the browser, no server needed, OpenAI API compatible.
Supports Llama 3, Phi 3, Gemma, Mistral, Qwen, and custom models in MLC format.
Offers streaming, JSON mode, and function calling; uses web workers/service workers.
Can be integrated into Chrome extensions; supports custom model integration.

Created 3 years ago

Updated 5 days ago

Starred 1 week ago by

Tomas Valenta(Cofounder of E2B). Also starred by

Gregor Zunic(Cofounder of Browser Use),

Travis Fischer(Founder of Agentic),

Taranjeet Singh(Cofounder of Mem0), and

3 more.

OpenSpec by Fission-AI

0.3%

70k

Spec-driven development for AI coding assistants

Aligns human and AI intent with structured specifications before code generation.
Integrates with popular AI coding assistants via slash commands or context rules.
Manages changes through proposals, tasks, and spec deltas for auditable development.
Enables deterministic AI outputs by locking requirements before implementation.

Created 1 year ago

Updated 1 day ago

Starred 2 weeks ago by

Dirk Englund(MIT EECS Professor and Cofounder of Axiomatic AI). Also starred by

Rodrigo Nader(Cofounder of Langflow).

crawlee-python by apify

0.1%

10k

Python library for web scraping and browser automation

Extracts data for AI, LLMs, RAG, or GPTs from websites.
Downloads HTML, PDF, JPG, PNG, and other files.
Uses BeautifulSoup, Playwright, and raw HTTP; headful/headless modes.
Includes proxy rotation, automatic retries, and session management.

Created 2 years ago

Updated 4 days ago

Starred 2 weeks ago by

Ralf Gommers(Core Contributor to scientific Python ecosystem). Also starred by

Clément Renault(Cofounder of Meilisearch),

Jiaming Song(Chief Scientist at Luma AI),

Wei-Lin Chiang(Cofounder of LMArena), and

4 more.

OpenBLAS by OpenMathLib

Optimized BLAS library for linear algebra subprograms

Optimizes performance with CPU-specific kernels (x86, ARM, Power, RISC-V, LoongArch).
Supports multiple operating systems including Linux, Windows, macOS, Android, and BSD.
Offers dynamic architecture support for runtime CPU detection and optimization.
Can be configured with OpenMP for multithreading.

Created 15 years ago

Updated 3 days ago

Starred 2 months ago by

Luis Capelo(Cofounder of Lightning AI).

compressed-tensors by vllm-project

326

Efficient safetensors extension for compressed LLM tensor storage

Stores sparse, quantized tensors on disk, unifying diverse compression formats.
Supports GPTQ, AWQ, INT8, FP8, SmoothQuant, and various sparsity patterns.
Integrates with PyTorch and Hugging Face for streamlined LLM deployment.
Enables flexible weight, activation, and KV cache quantization, including non-uniform schemes.

Created 2 years ago

Updated 1 day ago

Starred 2 months ago by

Elie Bursztein(Cybersecurity Lead at Google DeepMind). Also starred by

Shizhe Diao(Author of LMFlow; Research Scientist at NVIDIA).

Toolathlon by hkust-nlp

487

Benchmarking language agents for diverse, long-horizon tool use

Evaluates agents on 600+ real-world tools across realistic, complex, long-horizon tasks.
Utilizes Dockerized environments and decoupled loops for flexible agent scaffold integration.
Supports multiple agent frameworks including OpenAI SDK and OpenHands compatibility.
Provides a public evaluation service and trajectory visualization for analysis.

Created 11 months ago

Updated 1 month ago

Starred 2 months ago by