Projects Recently Starred by Experts

Starred 14 hours ago by

Tim Suchanek(Founder of expand.ai). Also starred by

Travis Fischer(Founder of Agentic),

Joe Walnes(Head of Experimental Projects at Stripe), and

Paul Stamatiou(Cofounder of Limitless).

tolaria by refactoringhq

8.7%

17k

Desktop app for managing markdown knowledge bases

Organize personal knowledge bases and company docs for AI context.
Files-first and Git-first principles ensure data portability and version control.
AI-first design integrates with agents, supporting local and offline use.
Built with Tauri, React, and TypeScript for a robust, open-source experience.

Created 4 months ago

Updated 10 hours ago

Starred 17 hours ago by

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

Jeff Hammerbacher(Cofounder of Cloudera) and

Amanpreet Singh(Cofounder of Contextual AI).

headroom by chopratejas

30.1%

30k

LLM context optimization layer

Compresses redundant LLM tool outputs and logs, achieving 47-92% token savings.
Optimizes context for code, logs, and JSON, preserving original data via reversible CCR.
Integrates with LangChain, Agno, and OpenAI clients via a zero-code-change proxy.
Enhances LLM provider caching and enables efficient multi-tool agent operations.

Created 5 months ago

Updated 9 hours ago

Starred 23 hours ago by

Luis Capelo(Cofounder of Lightning AI). Also starred by

Gregor Zunic(Cofounder of Browser Use).

CADAM by Adam-CAD

6.1%

Web app for AI-driven 3D CAD model creation

Transforms natural language and image inputs into 3D models using AI generation.
Offers interactive parametric controls, real-time previews, and exports to STL/SCAD.
Runs in-browser via OpenSCAD WASM, powered by a Supabase backend.
Integrates BOSL/MCAD libraries and supports custom fonts for enhanced design.

Created 9 months ago

Updated 1 day ago

Starred 1 day ago by

David Cournapeau(Author of scikit-learn).

VibeThinker by WeiboAI

46.6%

851

Small model, big logic: diversity-driven optimization for advanced reasoning

Achieves state-of-the-art reasoning in math and coding with only 1.5B parameters.
Uses innovative 'Spectrum-to-Signal Principle' (SSP) with diversity-exploring distillation and policy optimization.
Outperforms much larger models on challenging mathematical benchmarks like AIME and HMMT.
Offers significant cost savings in development compared to larger, state-of-the-art models.

Created 7 months ago

Updated 1 day ago

Starred 1 day ago by

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

OM1 by OpenMind

0.6%

Modular AI runtime for robots

Enables multimodal AI agents for digital environments and physical robots like humanoids and quadrupeds.
Processes diverse inputs (web, social media, camera, LIDAR) and enables physical actions (motion, navigation, conversation).
Supports hardware integration via plugins for ROS2, Zenoh, and CycloneDDS, with recommended Zenoh development.
Features pre-configured endpoints for Voice-to-Speech, GPT-4o, DeepSeek, and multiple VLMs.

Created 1 year ago

Updated 17 hours ago

Starred 1 day ago by

Pawel Garbacki(Cofounder of Fireworks AI). Also starred by

Chaoyu Yang(Founder of Bento) and

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

semantic-router by vllm-project

2.5%

Intelligent LLM routing for efficient inference

Routes requests to the best-suited model using semantic understanding and BERT classification.
Selects relevant tools based on prompt intent, reducing tokens and improving accuracy.
Detects and filters PII and jailbreak prompts for enhanced security and privacy.
Implemented in Go (with Rust FFI) and Python, with similarity caching for reduced latency.

Created 9 months ago

Updated 12 hours ago

Starred 1 day ago by

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

Jeff Hammerbacher(Cofounder of Cloudera).

GLM-5 by zai-org

9.8%

Large language model for complex agentic systems

Powers complex systems engineering and long-horizon agentic tasks with advanced reasoning and planning.
Leverages DeepSeek Sparse Attention for efficient long-context handling and reduced deployment costs.
Utilizes novel asynchronous RL infrastructure ('slime') for enhanced training efficiency and fine-grained post-training.
Achieves top open-source performance in coding, reasoning, and long-term operational agentic benchmarks.

Created 4 months ago

Updated 1 day ago

Starred 1 day ago by

Elie Bursztein(Cybersecurity Lead at Google DeepMind).

SkillOpt by microsoft

40.7%

Text-space optimization for LLM agent skills

Trains reusable natural-language skills for frozen LLM agents via trajectory optimization.
Mimics neural network training with epochs, batch sizes, and validation gates.
Leverages trajectory-driven edits and validation-gated updates for skill refinement.
Generates deployable best_skill.md artifacts and includes an optional monitoring WebUI.

Created 1 month ago

Updated 1 day ago

Starred 1 day ago by

Didier Lopes(Founder of OpenBB). Also starred by

Elie Bursztein(Cybersecurity Lead at Google DeepMind).

OpenKB by VectifyAI

1.8%

Open-source LLM knowledge base system

Compiles diverse documents into an interlinked, LLM-managed wiki for persistent knowledge accumulation.
Features vectorless long document retrieval via PageIndex and native multi-modal understanding (text, images, tables).
Leverages LLMs for knowledge synthesis, cross-referencing, and uses LiteLLM for multi-provider LLM support.
Outputs Obsidian-compatible Markdown, enabling graph visualization and knowledge exploration.

Created 2 months ago

Updated 12 hours ago

Starred 1 day ago by

Luis Capelo(Cofounder of Lightning AI). Also starred by

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