[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-843":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":10,"archived":22,"fork":22,"defaultBranch":23,"hasWiki":24,"hasPages":22,"topics":25,"createdAt":10,"pushedAt":10,"updatedAt":29,"readmeContent":30,"aiSummary":31,"trendingCount":16,"starSnapshotCount":16,"syncStatus":32,"lastSyncTime":33,"discoverSource":34},843,"USB-Uncensored-LLM","techjarves\u002FUSB-Uncensored-LLM","techjarves","The ultimate zero-install, portable local AI environment. Run high-quality, uncensored LLMs (Gemma, Qwen, NemoMix) directly from any USB drive or SSD. Fully air-gapped, cross-platform (Win\u002FMac\u002FLinux), and privacy-first with persistent chat history.","",null,"HTML",1476,311,31,25,0,48,104,548,144,102.48,false,"main",true,[26,27,28],"local-llm","portable-ai","uncensored-ai","2026-06-12 04:00:06","# USB-Uncensored-LLM ⚡\n\n**USB-Uncensored-LLM** is a fully air-gapped, zero-dependency, plug-and-play Local AI environment designed to run seamlessly from your **local hard drive** or a **portable USB\u002FSSD**. It bypasses complex installations natively executing large language models directly on your hardware with no internet required.\n\nWith a unified architecture, you can initialize your AI models once and choose to keep them on your system or carry them with you across Windows, macOS, and Linux PCs.\n\n🎥 **Watch the Setup & Demo Video:** [https:\u002F\u002Fyoutu.be\u002F60PSXsoXc8A](https:\u002F\u002Fyoutu.be\u002F60PSXsoXc8A)\n\n[![USB-Uncensored-LLM Setup & Demo](https:\u002F\u002Fimg.youtube.com\u002Fvi\u002F60PSXsoXc8A\u002Fmaxresdefault.jpg)](https:\u002F\u002Fyoutu.be\u002F60PSXsoXc8A)\n\n## 🚀 Core Features\n* **Zero Dependency Setup:** Ships with portable Python and isolated engine binaries. No system permissions, registry edits, or package managers required.\n* **Cross-Platform Interoperability:** Uses a intelligent `Shared` volume system — download your 5GB+ AI models *once*, and use them natively on Windows, macOS, and Linux without duplication.\n* **Censorship Free:** Integrates cutting-edge ablative and heretic fine-tuned models for completely unfiltered interactions.\n* **Network Proxied UI:** The custom Python HTTP server instantly serves a blazing-fast dark mode UI. You can access the AI from your phone or tablet on the same WiFi network without complex CORS configuration.\n* **Hardware Accelerated:** Uses a custom-compiled Ollama engine under the hood, natively capitalizing on AVX CPU instructions, NVIDIA CUDA, or Apple Metal GPU accelerators dynamically when plugged into different host machines.\n\n---\n\n## 💻 System Requirements\nBefore preparing your drive, ensure you have:\n- **Storage:** A USB 3.0+ flash drive or SSD with an absolute minimum of **8 GB** free space (16 GB is **highly** recommended).\n- **RAM:** The host computer should have at least **8 GB of system memory** to run the 2B\u002F4B models, and **16 GB of memory** to fluidly run the 9B\u002F12B models.\n\n---\n\n## 📂 Folder Architecture\n\nThe project is structured to strictly isolate operating system executables while securely unifying heavy model weights to save precious portable storage capacity.\n\n```text\n[Portable USB Drive]\n ├── 📁 Android    # Native Android (Termux) installers & launchers\n ├── 📁 Linux      # Native Ubuntu\u002FDebian offline installers & launchers\n ├── 📁 Mac        # Native macOS offline installers & launchers\n ├── 📁 Windows    # Native Windows offline automatic UI menus\n └── 📁 Shared     # Unified Data System\n      ├── 📁 bin         (Holds isolated executables: ollama-windows.exe, ollama-darwin...)\n      ├── 📁 chat_data   (Houses cross-platform persistent conversation history)\n      ├── 📁 models      (HuggingFace GGUF Weights & local database mapping)\n      └── 📁 python      (Isolated portable python environment)\n```\n\n---\n\n## 🧠 Curated AI Model Library\n\nThis USB ships with a curated installer for the highest-quality, locally operable uncensored models available on the open-source market today:\n\n1. **Gemma 2 2B Abliterated (~1.6 GB)**: *Recommended for all.* Extremely fast, incredibly smart for its size, with safety alignment vectors mathematically purged.\n2. **Gemma 4 E4B Ultra Uncensored Heretic (~5.34 GB)**: A \"heretic\" fine-tune that aggressively forces compliance to all user queries regardless of content or legality.\n3. **Qwen 3.5 9B Uncensored Aggressive (~5.2 GB)**: A much larger, incredibly competent reasoning model with a strict adherence to raw, unbiased answers.\n4. **Custom Models**: The installer supports downloading *any* .gguf weight directly from HuggingFace natively into the USB's engine.\n\n---\n\n## ⚙️ Quick Start Guide\n\n### Step 1: Initialize the Engine\nDepending on the computer you are currently plugged into, navigate into the respective Operating System folder and double-click\u002Frun the install script. \n* **Windows:** Double-click `Windows\u002Finstall.bat`\n* **macOS:** Open Terminal, drag in `Mac\u002Finstall.command`, and press Enter.\n* **Linux:** Run `bash Linux\u002Finstall.sh`\n* **Android:** Open Termux, run `bash Android\u002Finstall.sh` (see Android section below)\n\n> **Note:** Initializing simply downloads the tiny 50MB execution engine specific to that computer to the `Shared\u002Fbin` folder. \n\n### Step 2: Download AI Models \nIt is highly recommended to run the model download phase via a **Windows PC** (`Windows\u002Finstall.bat`), which provides an interactive, terminal-based catalog to easily select and download highly curated, uncensored GGUF Models. \n*(If you do not have a Windows PC, simply download your `.gguf` weights from HuggingFace and place them into the `Shared\u002Fmodels` folder manually).*\n\n### Step 3: Launch\nOpen the respective OS folder and run the `start` script:\n* **Windows:** `Windows\u002Fstart-fast-chat.bat`\n* **macOS:** `Mac\u002Fstart.command`\n* **Linux:** `bash Linux\u002Fstart.sh`\n* **Android:** `bash Android\u002Fstart.sh` (in Termux)\n\nThe engine will spin up securely in the background, and your default web browser will automatically open the locally-served Chat UI.\n\n---\n\n## 🏠 Local Disk Installation\nWhile this project is optimized for USB portability, it works beautifully as a lightweight local AI setup on your primary computer.\n\n**How to Install Locally:**\n1.  **Download\u002FClone** this repository to a folder on your `C:\\` or `D:\\` drive.\n2.  Navigate to the **Windows** (or Mac\u002FLinux) folder.\n3.  Run **`install.bat`** and choose your desired models.\n4.  The system will download everything into that local folder. \n5.  Run **`start-fast-chat.bat`** to begin.\n\n*Benefit:* Running from an internal SSD is significantly faster than a USB drive, resulting in near-instant AI model loading!\n\n---\n\n## 📱 Android Native (Termux)\nRun the AI engine **directly on your Android phone or tablet** — no PC required!\n\n### Requirements\n- **Termux** installed from [F-Droid](https:\u002F\u002Ff-droid.org\u002Fen\u002Fpackages\u002Fcom.termux\u002F) (NOT the Play Store — it's outdated)\n- **6 GB+ RAM** (8 GB+ recommended). Only the 2B model runs well on 6 GB devices.\n- **WiFi or mobile data** for initial setup (downloading engine + models)\n- **ARM64 processor** (virtually all modern Android phones\u002Ftablets)\n\n### Setup\n1. Copy the USB-Uncensored-LLM folder to your Android device (via USB OTG, file transfer, or `git clone`)\n2. Open **Termux** and navigate to the project folder\n3. Run: `bash Android\u002Finstall.sh`\n4. Select your model (Gemma 2 2B recommended for most Android devices)\n5. Wait for downloads to complete — **keep Termux in the foreground!**\n\n### Launch\n```bash\nbash Android\u002Fstart.sh\n```\nThe AI engine starts and Chrome opens automatically with the chat UI.\n\n### Android Performance Tips\n- **Run `termux-wake-lock`** before starting — prevents Android from killing the process\n- **Keep Termux in the foreground** for best performance\n- **Close other apps** to free RAM for the AI model\n- **Use the 2B model** on devices with less than 12 GB RAM\n- **Plug in your charger** — LLM inference drains battery fast\n- Expect **~3-10 tokens\u002Fsec** on the 2B model (vs 30-50+ on a PC with GPU)\n\n---\n\n## 📱 LAN Mobile Access\nIf you want to use the Heavyweight AI from your phone while lounging on the couch:\n1. Ensure your PC running the `start` script and your phone are on the exact same WiFi network.\n2. The terminal window will automatically detect your host machine and display a **Network Access** IP Address (e.g., `http:\u002F\u002F192.168.1.15:3333`).\n3. Simply type that URL into your mobile browser (Safari\u002FChrome). The custom Python server perfectly routes mobile queries directly to the USB! *(Note: If pages do not load, ensure Windows Firewall allows incoming connections on port `3333`)*.\n\n---\n\n## 🛠️ Troubleshooting\n\n- **The script instantly closes on Windows:** You likely have the legacy Windows App Execution Aliases turned on, which tricks the OS. Run the script via a command prompt, or right-click the `.bat` file and \"Run as Administrator\".\n- **\"Ollama Engine Not Found\":** You attempted to run the `start` script before the `install` script downloaded the base software for your specific OS. Run your OS's installer!\n- **Slow Generation Speeds:** Your model is too large for your host PC's RAM. Re-run `install.bat` and select the **Gemma 2 2B Abliterated** model, which runs rapidly even on older machines.\n\n---\n> *Disclaimer: USB-Uncensored-LLM is built for uncompromising computational freedom. By utilizing ablative models, the system will not moralize, lecture, or refuse your prompts. Please use responsibly.*\n# USB-Uncensored-LLM\n","USB-Uncensored-LLM 是一个零安装、便携式的本地AI环境，可以从任何USB驱动器或SSD直接运行高质量、无审查的大语言模型（如Gemma、Qwen、NemoMix）。其核心功能包括无需依赖的即插即用设置、跨平台兼容性（支持Windows、macOS和Linux）、完全无审查的内容生成以及硬件加速支持。该项目特别适合需要在不同设备间移动使用AI模型且重视隐私保护的场景，如开发者、研究人员或个人用户。通过内置的Python HTTP服务器，用户可以在同一WiFi网络下的手机或平板上访问AI界面，无需复杂的CORS配置。",2,"2026-06-11 02:39:46","CREATED_QUERY"]