Projects Recently Starred by Experts

Starred 2 hours ago by

Taranjeet Singh(Cofounder of Mem0). Also starred by

Tobi Lutke(Cofounder of Shopify),

Didier Lopes(Founder of OpenBB),

John Philip Morgan(Cofounder of Jasper), and

9 more.

qmd by tobi

14.6%

AI-powered local search for your documents

Indexes markdown, notes, and docs using BM25 and vector embeddings.
Leverages LLM re-ranking and query expansion for precise results.
Facilitates agentic workflows with structured output and MCP server.
Runs entirely locally via Ollama, ensuring data privacy.

Created 2 months ago

Updated 1 day ago

Starred 20 hours ago by

Amanpreet Singh(Cofounder of Contextual AI). Also starred by

Dan Guido(Cofounder of Trail of Bits).

nightshift by marcus

3916.7%

348

AI-powered nightly code maintenance

Finds and fixes code decay, doc drift, and security gaps via automated PRs.
Leverages Claude and Codex LLM budgets for cost-effective, overnight code analysis.
Acts as a 'code Roomba,' creating PRs that can be merged or closed.
Supports AI agent skill grooming and integrates with LLM provider CLIs.

Created 1 week ago

Updated 15 hours ago

Starred 21 hours ago by

Elie Bursztein(Cybersecurity Lead at Google DeepMind).

code-mode by universal-tool-calling-protocol

0.6%

Code execution library for AI agents, optimizing tool calls

Enables agents to execute complex toolchains as single TypeScript code blocks, drastically improving efficiency.
Supports universal protocols (MCP, HTTP, File, CLI) with secure Node.js VM sandboxing for robust execution.
Achieves significant speedups, token reduction, and cost savings versus traditional multi-call tool orchestration.
Features dynamic tool discovery, auto-generated interfaces, and built-in observability for seamless agent integration.

Created 3 months ago

Updated 1 week ago

Starred 23 hours ago by

Vincent Weisser(Cofounder of Prime Intellect).

AgentLab by ServiceNow

0.2%

513

Open-source framework for web agent development, testing, and benchmarking

Scalable agent experiments using Ray, targeting BrowserGym benchmarks.
Unified LLM API supports OpenRouter, OpenAI, Azure, and self-hosted TGI.
Reproducibility features track software versions and manage stochasticity.
AgentXray provides a Gradio interface for visualizing agent traces.

Created 1 year ago

Updated 1 week ago

Starred 23 hours ago by

Vincent Weisser(Cofounder of Prime Intellect). Also starred by

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

Gregor Zunic(Cofounder of Browser Use), and

Jiayi Pan(Author of SWE-Gym; MTS at xAI).

BrowserGym by ServiceNow

0.3%

Gym environment for web task automation research

Web agent training/evaluation across MiniWoB, WebArena, VisualWebArena, WorkArena, AssistantBench.
Supports open-ended tasks via interactive chat with agents; OpenAI API demo agent.
Uses Playwright; benchmarks require inheriting AbstractBrowserTask class.
AgentLab framework for implementing, testing, and evaluating web agents.

Created 2 years ago

Updated 1 week ago

Starred 1 day ago by

Joel Griffith(Founder of Browserless). Also starred by

PJ Hyett(Cofounder of GitHub),

Shyamal Anadkat(Research Scientist at OpenAI),

Tim J. Baek(Founder of Open WebUI), and

39 more.

openclaw by openclaw

11.3%

202k

WhatsApp automation and AI integration for machine actions

Send, receive, and auto-reply to WhatsApp messages using Twilio or personal WhatsApp Web sessions.
Integrates with AI models like Claude via CLI, supports text, media, and voice note transcription.
One-command webhook setup with Tailscale Funnel for seamless inbound message handling.
Enables personal AI assistants and machine control triggered by WhatsApp messages.

Created 2 months ago

Updated 12 hours ago

Starred 1 day ago by

Travis Fischer(Founder of Agentic).

ClawRouter by BlockRunAI

29.2%

Smart LLM routing for cost optimization

Automatically routes requests to the cheapest capable LLM, saving up to 78% on inference costs.
Employs 14-dimension weighted scoring for intelligent, 100% local routing decisions in under 1ms.
Integrates x402 micropayments via USDC on Base, enabling non-custodial, per-request payments for agents.
Supports over 30 models from major providers through a single, auto-generated wallet without API keys.

Created 2 weeks ago

Updated 18 hours ago

Starred 1 day ago by

Jason Miller(Author of Preact).

sled by layercodedev

4.6%

270

Access coding agents from your phone with voice

Control local Claude Code, Codex, or Gemini CLI agents via a mobile web UI.
Enables voice input and output for hands-free interaction with coding assistants.
Runs entirely locally, keeping code and prompts private on your machine.
Leverages ACP for agent communication and Hono/Cloudflare Workers for the web UI.

Created 1 month ago

Updated 3 weeks ago

Starred 1 day ago by

Johannes Schickling(Cofounder of Prisma). Also starred by

Pawel Garbacki(Cofounder of Fireworks AI) and

Beyang Liu(Cofounder of Sourcegraph).

butterfish by bakks

1.0%

496

CLI tool for adding AI to your shell

Adds AI prompting to your shell (bash, zsh) using OpenAI.
Autocompletes shell commands and answers contextual questions using shell history.
Goal mode enables agentic workflows, executing commands to achieve objectives.
Supports local models via OpenAI-compatible API, file summarization, and embedding index.

Created 3 years ago

Updated 1 week ago

Starred 1 day ago by

Dan Guido(Cofounder of Trail of Bits). Also starred by

Alexey Milovidov(Cofounder of Clickhouse).

heretic by p-e-w

29.3%

Automatic LLM censorship removal

Decensors transformer language models via automated directional ablation and parameter optimization.
Supports multimodal and MoE architectures through a user-friendly CLI tool.
Minimizes model refusals while preserving original intelligence with low KL divergence.
Utilizes float-based direction interpolation for enhanced ablation parameter discovery.

Created 4 months ago

Updated 13 hours ago

...