Updated Sep 17, 2026 · rebuilt daily from GitHub data Every hub below pairs an editorial how-to-choose guide with a live ranking rebuilt each day from GitHub data and our own star snapshots. AI Image & Design This is the open-source side of AI image generation and design: diffusion models you can run yourself, ComfyUI-style pipelines, background removers, upscalers, and the toolkits that wrap them into something usable. Unlike the closed apps, these put the weights and the workflow in your hands — no per-image fees, no content gates. 88 repos ranked AI video is the fastest-moving corner of open-source AI right now: text-to-video models, talking-avatar and lip-sync tools, automatic subtitling and translation, and full pipelines that turn a prompt or a script into a finished clip. Most of the buzz is locked inside paid apps — this page tracks the open repos you can actually run and build on, ranked daily by real GitHub momentum so you see what's genuinely taking off versus what just launched loudly.. 92 repos ranked Agent frameworks are where LLMs stop being chatbots and start doing work: planning steps, calling tools, checking their own output, and looping until a task is done. The space moves brutally fast — last quarter's hot framework is this quarter's maintenance-mode repo — so star velocity and commit recency matter more here than almost anywhere else on GitHub. 112 repos ranked Everything around the model: inference servers, fine-tuning kits, RAG plumbing, eval harnesses, prompt tooling, and local runners. This is the fastest-compounding category on GitHub — projects here routinely gain thousands of stars in a week when they catch a wave. 98 repos ranked Developer Tools CLI utilities, editors and editor plugins, terminal upgrades, code search, formatters, and the general quality-of-life layer of software work. Dev tools age differently from frameworks: the good ones get adopted into muscle memory and stay there for a decade. 129 repos ranked Frameworks, meta-frameworks, UI libraries, and build tooling for the web. Star counts here are the most inflated on GitHub — everything popular is already huge — so the interesting signal is the second derivative: which tools are still accelerating versus coasting on installed base. 81 repos ranked Software you run on your own box instead of renting: media servers, photo libraries, password managers, analytics, document stores, dashboards, and whole-SaaS replacements. Self-hosted projects earn stars slowly and keep them forever — the category rewards reliability over hype. 139 repos ranked Data Engineering Pipelines, orchestrators, stream processors, dataframe engines, and the connective tissue between databases and decisions. Data tooling has consolidated hard around a few winners, which makes the newcomers that do break through worth a serious look — they usually solve a real pain (cost, speed, or developer experience) rather than rebranding an old idea. 82 repos ranked Relational engines, vector stores, key-value caches, time-series systems, and embedded databases. A database choice outlives every other dependency in your stack, so the bar for switching is high — and yet this category keeps producing genuine breakouts, especially where AI workloads meet storage.

82 repos ranked Scanners, secret detectors, dependency auditors, network analysis, and offensive-security tooling for authorized testing. Security tools on GitHub split into two populations: professional-grade projects with steady maintenance, and proof-of-concept drops that spike and stall. 83 repos ranked Containers, Kubernetes ecosystem, CI/CD, infrastructure-as-code, observability, and the machinery that keeps services running. This category rewards depth: the projects that win are the ones operators trust at 3 a.m. 83 repos ranked The best open-source web scrapers on GitHub — the tools that turn any website into clean, structured data. Crawlers, browser automation, anti-bot evasion for legitimate data collection, HTML parsers, and the new wave of LLM-powered extractors that turn pages into clean JSON. 87 repos ranked Productivity & Knowledge Note-taking systems, todo managers, calendars, launchers, and personal knowledge bases — the open-source answers to a very crowded SaaS aisle. Open-source productivity tools win on three things commercial ones can't match: data ownership, plain-text portability, and plugin ecosystems. 93 repos ranked Learning & Interview Prep Roadmaps, course repositories, algorithm collections, system-design guides, and build-your-own-X tutorials. These are some of the most-starred repos on all of GitHub because they serve every developer alive — which also means quality varies wildly within the same star bracket. 89 repos ranked Local & Private AI The open-source stack for running AI yourself: local model runners, private ChatGPT-style UIs, on-device RAG, and self-hosted agents that never send your data to a third party. As model weights get smaller and laptops get faster, running capable AI locally went from a hobby to a real alternative to paid APIs. 83 repos ranked Papers with Code The bridge between AI research and working software: open-source implementations of machine-learning papers, official model releases, and reproducible research code. Since the original Papers with Code wound down, there is no single live place to find which paper implementations developers are actually adopting. 81 repos ranked Computer Vision Computer vision is the part of AI that reads the physical world: detecting objects, segmenting images, recognizing faces and text, estimating human pose, and following movement across video. The field ships fast on GitHub, from production detectors like the YOLO family to research code for segmentation, depth estimation, and SLAM. 80 repos ranked