Links
AddLet Claude (or any LLM) actually watch a video — scene-aware, deduplicated frames + transcript, from a URL or local file. Runs locally, MIT.
The local control plane for long-running AI agent work.
A lightweight state kernel and agent-agnostic local control plane for loop engineering, LoopX keeps long-running work reviewable, restartable, and easier to hand off across turns, tools, and agents. It does not replace your agent runtime.
Lightweight loop engineering state kernel for long-running AI agent teams. Agent-loop agnostic across Codex, Claude Code, and other coding agents, with durable goals, quota-aware auto-wake, executable todos, evidence logs, and verifiable handoffs.
Revisiting Diffusion Transformers for Monocular Depth Estimation.
Marigold V2 is a family of models and a cost-effective fine-tuning protocol that repurposes a pretrained diffusion transformer into single-step dense predictors: depth, see-through depth, surface normals, albedo, and other dense modalities. Fine-tuning takes less than a week on a single consumer GPU, within reach of individual practitioners and small labs, and the results are state of the art, faithfully reproducing sharp edges, fur, and hair-thin details. The same models also unlock applications such as metric depth completion.
Self hosted, easy to install end to end encrypted storage drive.
Hoodik is a lightweight, secure, and self-hosted cloud storage solution. It's designed and built with Rust and Vue, focusing on end-to-end encryption that shields your data from prying eyes and hackers. Hoodik supports file uploading and downloading, making it easy for you to share files with other users. The simple and intuitive web interface makes file management a breeze. Thanks to Rust's focus on speed and performance, your data transfers will be lightning fast.
Bypass DPI. Implementation of some DPI bypass methods. The program is a local SOCKS proxy server.
Related contents:
Distilled variant of Whisper for speech recognition. 6x faster, 50% smaller, within 1% word error rate.
Durable, searchable memory of your past agent sessions.
Durable memory for your AI coding agents. funes indexes your past sessions across Claude Code, Codex, pi, and Hermes and lets any agent recall the past decisions, rationale, and findings. Your memory is a dataset you can publish to the Hugging Face Hub — then any machine, teammate, or agent can recall from it.
Related contents:
🤗 ml-intern: an open-source ML engineer that reads papers, trains models, and ships ML models
An ML intern that autonomously researches, writes, and ships good quality ML related code using the Hugging Face ecosystem — with deep access to docs, papers, datasets, and cloud compute.
🔥The Web-scale GUI for MongoDB.
MongoDB client for the web. Query your data directly from your browser. You can host it locally, or anywhere else, for you and your team.
🤗 smolagents: a barebones library for agents. Agents write python code to call tools and orchestrate other agents.
A novel-writing tool created by a novelist. A distraction-free word processor for authors, by a wannabe author.
NEO understands from the moment you install it that you are writing books and nothing else. No bloat, no distractions, with manuscripts that look like books as you write them.
Create agents that monitor and act on your behalf. Your agents are standing by!
Huginn is a system for building agents that perform automated tasks for you online. They can read the web, watch for events, and take actions on your behalf. Huginn's Agents create and consume events, propagating them along a directed graph. Think of it as a hackable version of IFTTT or Zapier on your own server. You always know who has your data. You do.
Principles for building reliable LLM applications.
What are the principles we can use to build LLM-powered software that is actually good enough to put in the hands of production customers?
A Modern And Secure CLI Tool For Managing Environment Variables.
Envio is a command-line tool that simplifies the management of environment variables across multiple profiles. It allows users to easily switch between different configurations and apply them to their current environment.
Useful tool to track location or mobile number.
Machinae is a tool for collecting intelligence from public sites/feeds about various security-related pieces of data: IP addresses, domain names, URLs, email addresses, file hashes and SSL fingerprints. It was inspired by Automater, another excellent tool for collecting information. The Machinae project was born from wishing to improve Automater in 4 areas:
A terminal-based AWS resource viewer and manager.
taws provides a terminal UI to interact with your AWS resources. The aim of this project is to make it easier to navigate, observe, and manage your AWS infrastructure in the wild. taws continually watches AWS for changes and offers subsequent commands to interact with your observed resources.
Using an LLM to call tools in a loop is the simplest form of an agent. This architecture, however, can yield agents that are “shallow” and fail to plan and act over longer, more complex tasks. Applications like “Deep Research”, "Manus", and “Claude Code” have gotten around this limitation by implementing a combination of four things: a planning tool, sub agents, access to a file system, and a detailed prompt.
deepagents is a Python package that implements these in a general purpose way so that you can easily create a Deep Agent for your application.
GUI file synchronization client that can sync with any cloud provider. Backed by rclone, giving you a reliable and battle-tested way to sync your files anywhere.
Rust implementation of the H3 geospatial indexing system.
Related contents:
Use any icon (100,000+) from Iconify, for TailwindCSS.
A dynamic windowed scrolling table component for react.
HighTable is a virtualized table component for React, designed to efficiently display large datasets in the browser. It loads and renders only the rows necessary for the current viewport, enabling smooth scrolling and performance even with millions of rows. HighTable supports asynchronous data fetching, dynamic loading, and optional column sorting.
Related contents:
JavaScript Iceberg Client.
Icebird is a library for reading Apache Iceberg tables in JavaScript. It is built on top of hyparquet for reading the underlying parquet files.
Hyperlight is a lightweight Virtual Machine Manager (VMM) designed to be embedded within applications. It enables safe execution of untrusted code within micro virtual machines with very low latency and minimal overhead.
Hypothetical Indexes for PostgreSQL. HypoPG is a PostgreSQL extension adding support for hypothetical indexes.
An hypothetical -- or virtual -- index is an index that doesn't really exists, and thus doesn't cost CPU, disk or any resource to create. They're useful to know if specific indexes can increase performance for problematic queries, since you can know if PostgreSQL will use these indexes or not without having to spend resources to create them.
A terminal image and video viewer.
A user-friendly terminal image viewer that uses graphic capabilities of terminals (Sixel, Kitty or iterm2), or 24-Bit color capabilities and unicode character blocks if these are not available.
On terminals that implement the sixel protocol, the Kitty Graphics Protocol, or the iTerm2 Graphics Protocol this displays images in full resolution.
Fast, stateless LLM-powered assistant for your shell: qq answers; qa runs commands.
qq means quick question. qa means quick agent. Both are easy to type rapidly on QWERTY keyboards with minimal finger movement. That makes interacting with LLMs faster and more natural during real work.
Stop logging into a black hole. Give your logs a beautiful, powerful, and interactive home.
log-vwer is the ultimate plug-and-play toolkit for Node.js developers who are tired of messy console.log statements. It instantly gives you a professional-grade dashboard to view, search, and filter your application logs in real-time.
Peergos is a peer-to-peer encrypted filesystem with secure sharing of files designed to be resistant to surveillance of data content or friendship graphs. It will have a secure email replacement, with some interoperability with email. There will also be a totally private and secure social network, where users are in control of who sees what (executed cryptographically).
50 open-source skills that teach your AI coding agent to make motion graphics, animation & video — kinetic typography, data-viz, explainers, TikTok/Reels, WebGL, Manim. 14 installable packs. By iart.ai, the AI motion agent.
Real-time router/firewall traffic visualizer TUI for Linux (interfaces, conntrack, firewall drops).
Where a host monitor shows your machine's own traffic, fwtop is built for the box that sits between networks — an edge router or firewall. It reads kernel state directly (near-zero overhead, no packet capture) to show per-interface throughput (split into WAN/LAN zones), the live connection-tracking table (including NAT), and firewall rule counters with a smoke→fire drops heatmap that surfaces spikes at a glance.
Related contents:
Multitouch gestures with libinput driver on Linux. Fusuma is multitouch gesture recognizer. This gem makes your linux able to recognize swipes or pinchs and assign commands to them.
🔍 OSINT tool for searching people's digital footprint and leaked passwords across various social networks, written in Go.
Related contents:
A fast and ergonomic concurrent hash-table for read-heavy workloads.
An Active Directory security-assessment toolkit in Rust: a PingCastle-class auditor that maps a domain's attack paths — scored, graphed, and MITRE-tagged — then, for authorized red-team and research use, proves those paths end-to-end. One static binary, from Kali/Linux or Windows, on an embedded from-scratch DCE/RPC · NTLM · SMB2 · Kerberos stack (the "impacket for Rust" that didn't otherwise exist).
A cross-platform GUI library for Rust, inspired by Elm. A cross-platform GUI library for Rust focused on simplicity and type-safety.
Use LLMs and LLM Vision (OCR) to handle paperless-ngx - Document Digitalization powered by AI.
paperless-gpt seamlessly pairs with paperless-ngx to generate AI-powered document titles and tags, saving you hours of manual sorting. While other tools may offer AI chat features, paperless-gpt stands out by supercharging OCR with LLMs—ensuring high accuracy, even with tricky scans. If you’re craving next-level text extraction and effortless document organization, this is your solution.
Create beautiful scrolling driven navigation lists with stacking headers that remain visible at all times. Scroll around on the plugin homepage to see it in action.
Ballyregan is a package & CLI that allows you to fetch free tested proxies really fast!
Super-scale your images and run experiments with Residual Dense and Adversarial Networks. The goal of this project is to upscale and improve the quality of low resolution images. This project contains Keras implementations of different Residual Dense Networks for Single Image Super-Resolution (ISR) as well as scripts to train these networks using content and adversarial loss components.
A simple CLI tool to retrieve S3 storage costs and storage tiers for buckets in your AWS account.
RE# - A high-performance, automata based regex engine with first-class support for intersection and complement operations.
RE# compiles patterns into deterministic automata. All matching is non-backtracking with guaranteed linear-time execution. RE# extends System.Text.RegularExpressions syntax with intersection (&), complement (~), and a universal wildcard (_), enabling patterns that are impossible or impractical to express with standard regex.
Related contents:
Pretty simple jQuery plugin that turns standard Bootstrap alerts into "Growl-like" notifications.
🔮 Platform-neutral Browser Use for AI agents: real Chrome automation with a CLI + SDKs, no lock-in, dead simple.
open-browser-use is a browser automation layer that stays neutral across agent runtimes. It is also an open-source alternative to the Chrome Browser Use capability recently shipped in Codex.app. For the story behind it, see the Browser Use Deep Dive.
Related contents:
open-computer-use is an open-source Computer Use service wrapped as MCP. Any AI agent or MCP client can use it to run Computer Use on macOS, Linux, and Windows.
Related contents:
This repository will contain many mindmaps for cyber security technologies, methodologies, courses, and certifications in a tree structure to give brief details about them.
DirPlayer is a Shockwave Player emulator written in Rust that aims to make playing old browser games possible on modern browsers.
Related contents:
<html> as a Response.
mono-jsx is a JSX runtime that renders <html> element to Response object in JavaScript runtimes like Node.js, Deno, Bun, Cloudflare Workers, etc.
RSS Youtube Podcast Feed Generator.
This is a GO application that will take any podcast that is on Youtube and will generate a RSS feed with the audio only and all sponsored sections auto removed. The actual podcasts episodes are downloaded on demand from youtube when the user requests the specific episode then it is served to the user seamlessly.
A lightweight Material YouTube Music player for Android, built with Kotlin and Jetpack Compose.
Designed for simplicity and smooth performance, Umihi Music lets you enjoy your favorite tracks effortlessly.
Related contents:
Local Video Knowledge Base.
Local-first AI video intelligence platform. Index your video library with multi-modal analysis (YOLO, DeepFace, Whisper), search semantically via natural language, Docker-ready.
Edit Mind lets you index your videos (including transcription, frame analysis, and multi-model embedding), and you can search your videos (or specific video scenes) using natural language.
An open-source low-code Platform for Developers. ILLA is a robust open source low-code patform for developers to build internal tools. By using ILLA's library of Components and Actions developers can save massive amounts of time on building tools.
Simple, CLI-friendly secret storage that lets you safely commit encrypted secrets to version control.
Simple, password-based encrypted vault for .env and infrastructure secrets. Like git-crypt or sops, but dramatically simpler. Ideal for small teams and IaC workflows
lockenv provides a secure way to store sensitive files (like .env files, configuration files, certificates) in an encrypted .lockenv file that can be safely committed to your repository. Files are encrypted using a password-derived key and can be easily extracted when needed.
Linux GPU Configuration And Monitoring Tool.
This application allows you to control your AMD, Nvidia or Intel GPU on a Linux system.
Related contents:
AI by Hand ✍️ Exercises in Excel
Fearless refactoring, it does a lot of smart checks to find certain errors.