gitwtfhub

Popular Python repositories on GitHub, each decoded: what it actually does, its stack, and how to run it.

• django-haystack/pysolrPysolr is a lightweight Python library that lets your application talk to Apache Solr, a popular search server
• omnigent-ai/omnigentA common control layer over Claude Code, Codex, Pi, and custom agents that syncs sessions across devices and a
• facebookresearch/sapiens2A family of AI vision models from Meta, pretrained on a billion human images, for tasks like pose detection, s
• depthfirstdisclosures/nginx-riftA proof of concept exploit for a critical NGINX vulnerability that lets an attacker run code on a vulnerable s
• fluxions-ai/vuiVui is a self-hosted, real-time voice assistant that runs speech recognition, a local language model, and text
• karpathy/hn-time-capsulehn-time-capsule pulls a Hacker News front page from exactly ten years ago and uses AI to grade which commenter
• sophomoresty/gemini-web2apigemini-web2api is a single Python script that exposes Google Gemini through an OpenAI-compatible local API, us
• bakrianoo/mini-ragMini-RAG is a step-by-step video course with code that teaches you to build a real question-answering system o
• bytedance/lanceByteDance research release of a 3B-active-parameter unified multimodal model that handles image and video unde
• huangchihhungleo/claude-real-videoA local Python tool that extracts scene-change keyframes and audio transcripts from a video so you can paste t
• mattzh72/articraftArticraft uses AI to automatically generate 3D models of objects with moving parts, like hinges or sliders, by
• sapientinc/hrm-textPretraining stack for a 1B-parameter hierarchical recurrent language model that can be trained from scratch on
• qiuqiubuchongle-cloud/chokepoint-atlasA research tool that maps supply chains in AI-related industries to find which part will run out of capacity f
• lillian039/elfELF is a research codebase for a diffusion-based AI text generation model that refines text in a continuous em
• tencent-hunyuan/unirlTencent's reinforcement learning framework for training image, video, and language AI models with a shared tra
• facebookresearch/boxerBoxer turns 2D object detections in photos into 3D bounding boxes for indoor scenes, giving position, orientat
• facebookresearch/vicregVICReg trains image-understanding models using pairs of similar images instead of human labels, so you need fa
• khrisat/text-humanizerAn open source pipeline that rewrites AI-generated text through multiple translation passes to evade AI detect
• facebookresearch/vggt-omegaAn AI model that turns a set of photos or a video into 3D camera positions and depth, from Meta AI and Oxford.
• bytevisionlab/dreamliteA small AI model that generates and edits images directly on a phone in seconds, with no cloud needed.
• paddlepaddle/paddletsPaddleTS is a Python toolkit for building predictive models from time series data. It provides pre-built deep
• dsd2077/cyberverseAn open source platform for building real time voice and video AI agents and digital human characters using We
• truegunsmithfence/hentaihunterA command line tool for bulk downloading adult manga and doujinshi image galleries from specific hosting sites
• shang-zhu/violinAn open source tool that automatically dubs a video into another language, transcribing, translating, and re v
• baoweise-bot/aimili-vpngateA Python tool that gives a Linux server a rotating outbound IP by routing its traffic through free VPNGate ser
• automatic1111/stable-diffusion-webui-promptgenAn add-on for Stable Diffusion's web UI that uses fine-tuned GPT-2 models to auto-generate creative image prom
• kappaemme-git/codex-complexity-optimizerA Codex skill that scans a codebase for complexity and performance problems and writes a safe, actionable opti
• tianhangzhuzth/fundamental-avaA Python framework for simulating large populations of autonomous agents with memory, beliefs, and social mode
• juk1-gh/gpt-promo-scannerA Python tool that guesses and validates unpublished partner discount codes for ChatGPT Business by probing Op
• pluviobyte/video-production-skillsA growing library of video production skills for AI coding assistants like Claude Code and Codex, covering mot
• vast-ai-research/triposplatTurns a single photo into a 3D Gaussian Splatting file you can view or use in real time 3D tools.
• shandaai/alayarendererAn AI rendering research project that breaks game video into visual data layers and can then regenerate it wit
• xiaohuailabs/xiaohu-video-translateAI coding agent skills that auto-download, transcribe, translate, and burn Chinese subtitles onto foreign-lang
• joeseesun/qiaomu-goal-meta-skillA skill that turns a vague request into a structured Codex or Claude Code goal command with built in verificat
• patchfighterway90/cs2-external-overlayA Windows tool that reads game data externally to draw a customizable, click-through overlay on top of Counter
• kasothaphie/genreconA research system that reconstructs detailed 3D indoor scene meshes with surface materials from multiple photo
• ensignkazekage/prodigy-hacksA parent-companion tool for Prodigy Math Game that generates richer reports than the official dashboard, but i
• tiantiangpu/reg-factoryAn automated pipeline that registers batches of accounts on services like Gmail and ChatGPT using fingerprinti
• karpathy/researchpoolerA Python proof-of-concept that lets you search, filter, and bulk-open academic papers with a few lines of code
• facebookresearch/pixioA pixel-level vision model from Meta that excels at depth estimation, segmentation, and 3D reconstruction, bea
• ace-trump-tech/deltaforce-obs-lockerA game cheat tool for Delta Force that uses AI object detection to auto-aim at enemies and tries to avoid the
• chroma-core/context-1-data-genThe data generation pipeline behind Chroma's Context-1 search agent, producing synthetic multi-step search tra
• gordensun/gordenpptskillA Python skill that lets AI agents generate polished, Chinese-style PowerPoint decks from templates and a JSON
• poseljacob/agentic-video-editorA command line tool that uses a team of AI agents plus FFmpeg to turn raw video clips and a short brief into a
• flashml-org/flashlibFlashLib is a Python library offering GPU-accelerated classical machine learning operations like clustering, d
• scenemaai/scenema-audioA local voice cloning and expressive speech generation tool that performs emotional delivery, not just text-to
• facebookresearch/spinquantA technique that shrinks large language models to run faster and use less memory by rotating their data before
• 0xsero/codex-shimCodex-shim is a local Python server that lets Codex Desktop connect to any AI model provider, not just the app
• evilsocket/auditCLI security audit agent that runs Claude through an eight stage pipeline (Recon, Hunt, Validate, Gapfill, Ded
• jmmy9609-design/gpt-ppA gateway tool that intercepts Stripe checkout sessions to extract PayPal authorization links under a zero-amo
• facebookresearch/spdlA data loading library that makes building fast, scalable data pipelines for machine learning easier, processi
• karpathy/covid-sanityA searchable database of COVID-19 research papers that uses machine learning to find similar papers, helping r
• facebookresearch/actionmeshA tool that converts a short video of movement into an animated 3D mesh you can import into Blender or other 3
• fullive-ai/animaAnima is a local smart home operating system that connects your devices to an AI brain, letting them learn pre
• s-sigdel/vimhjklVimhjkl is a Vim practice tool with 222 keystroke-graded drills and spaced repetition that teach real editing
• hidream-ai/hidream-o1-imageAn open-source AI model that generates and edits images from text using one unified architecture instead of se
• novaquant-labs/hyperliquid-trading-botAn automated grid trading bot for Hyperliquid that places buy and sell orders at set intervals to profit from
• tong89/smartnodeSmartNode is a local simulation platform that visualizes how satellites, ground stations, and relay links coor
• juyterman1000/entrolyA proxy tool that trims what code gets sent to AI assistants to cut costs and checks AI answers against eviden
• jlevy/reprenA command-line tool that renames files and rewrites their contents at the same time using simultaneous, undo-s
• avaturn-live/avtr-1AVTR-1 turns a portrait photo and audio into a real-time lip-synced talking avatar video, with support for two
• zarazhangrui/personalized-podcastA Claude Code skill that turns articles, PDFs, or notes into a two-host audio podcast, published locally or sy
• kadevin/ilab-gpt-conjureA local web app for generating images with GPT-image-2, with a reference image library, prompt templates, task
• tencent-hunyuan/hy3d-benchA massive Tencent-released dataset of over 600,000 cleaned 3D objects, built to train and test AI models that
• hkust-c4g/anytalkerAn audio driven Python tool that turns a photo and a voice recording into a realistic talking head video, no c
• snazzybean/roommindA self learning Home Assistant integration that gives room by room climate control by modeling how each room h
• simonlin1212/tradingagents-astockA multi agent AI framework that analyzes Chinese A-share stocks using seven specialist agents and a bull versu
• gair-nlp/livetalkA real time research system that generates lip synced talking avatar video in under a third of a second per re
• txbabaxyz/polyrecA terminal dashboard that tracks Polymarket's 15-minute Bitcoin prediction markets in real time and helps back
• gainsec/autoproberAn AI agent controlled CNC rig that scans circuit boards with a microscope and safely probes chip pins for har
• evolink-ai/awesome-blender-seedance-workflow-usecasesA curated collection of 25 Blender and Seedance AI filmmaking workflows, showing how to use 3D previs and AI a
• yliust/tactileAn operating layer that helps AI agents control software by reading the accessibility tree instead of guessing
• tencent-hunyuan/hy-wuHY-WU is an open source AI research framework that edits images using text instructions, letting you transfer
• lynote-ai/humanize-textPython toolkit with four methods (translation chains, multi-turn LLM rewrites, detection-guided loops, mixed N
• agentic-in/elephant-agentElephant Agent is a personal AI assistant that builds a structured, correctable model of your identity, projec
• techwithtim/python-platformerA simple 2D side-scrolling platformer game in Python, built as a beginner learning project to show how games a
• kevrojo/dulusDulus is a free command-line AI coding agent that works by reusing your existing browser login sessions on ser
• adrienbrault/geekmagic-hacsThis Home Assistant integration turns a cheap GeekMagic mini display into a live smart home dashboard by rende
• hero8152/infinite-canvasA ComfyUI extension that lets you call external AI APIs and ModelScope models directly inside an infinite visu
• cslawyer1985/claude-for-legal-zhA set of Claude Code plugins that give Chinese lawyers AI help with contract review, litigation, corporate gov
• opendrivelab/simscaleA CVPR 2026 research project that improves self-driving car AI by mixing large-scale simulated driving scenari
• voicekit-team/t-oneA real time speech to text pipeline built specifically for Russian phone call audio, converting spoken convers
• chiennv2000/orthrusA technique that makes large language models generate text several times faster by running two coordinated vie
• paddlepaddle/interpretdlA toolkit that helps you understand why deep learning models make the decisions they do by highlighting which
• feder-cr/invisible_playwrightA patched, Playwright-compatible Firefox that spoofs browser fingerprint signals at the source-code level, aim
• minimax-ai/msaMSA is a GPU code library that speeds up the attention step inside AI language models on NVIDIA's newest chips
• pixel-talk/pearA research project that reconstructs a full 3D mesh of a person's body, face, and hands from a single photo or
• klotzkette/claude-fuer-deutsches-rechtCollection of Claude skills adapted for German legal practice, covering labor, corporate, insolvency, data pro
• mimic-video/mimic-videoA research project that trains robot control policies by attaching a small action decoder to a pretrained vide
• yangtiming/fast-sam-3d-bodyA research project that makes SAM 3D Body, a model that reconstructs 3D human poses from a single photo, run o
• lukashoel/video_to_worldA research codebase that reconstructs a consistent 3D scene from AI generated video by correcting the small 3D
• kyutai-labs/hibiki-zeroA real time AI model that translates spoken French, Spanish, Portuguese, or German into English speech, aiming
• jawaz-keyzor/darkgptA Python app that sends pre-written jailbreak prompts to the OpenAI API to try to bypass its content restricti
• hellogithub-team/hydraA Python web scraper that aggregates HelloGitHub's own content stats across Chinese platforms like WeChat and
• django-haystack/celery-haystackA Django tool that moves search index updates into the background using Celery, so saving or deleting content
• anybackup-ai/anybackupAn AI-agent-driven backup and recovery platform that turns plain-English requests into confirmed, logged datab
• konbakuyomu/smartsearchA command-line tool giving AI agents and terminal users one consistent interface to search the web, fetch page
• wubing2023/paperspineSkill suite for Codex and Claude Code that drives a structured academic-paper rewrite or build, producing a fu
• nvlabs/mobilitygenMobilityGen generates labeled training data for mobile robots by simulating them in realistic 3D environments,
• chubbyguan/chubbyskillsA personal collection of AI agent skills for transcribing videos, managing an Obsidian knowledge base, trackin

All repos · Explain any other repo →