[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-73823":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":16,"stars7d":16,"stars30d":17,"stars90d":16,"forks30d":16,"starsTrendScore":16,"compositeScore":18,"rankGlobal":10,"rankLanguage":10,"license":19,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":20,"hasPages":20,"topics":22,"createdAt":10,"pushedAt":10,"updatedAt":23,"readmeContent":24,"aiSummary":25,"trendingCount":16,"starSnapshotCount":16,"syncStatus":26,"lastSyncTime":27,"discoverSource":28},73823,"secret-llama","abi\u002Fsecret-llama","abi","Fully private LLM chatbot that runs entirely with a browser with no server needed. Supports Mistral and LLama 3.","",null,"TypeScript",2674,169,20,19,0,3,58.99,"Apache License 2.0",false,"main",[],"2026-06-12 04:01:11","# Secret Llama\n\n![secret llama](https:\u002F\u002Fgithub.com\u002Fabi\u002Fsecret-llama\u002Fassets\u002F23818\u002F0bf43a95-4fe5-4c53-87bc-b558e5c4968f)\n\nEntirely-in-browser, fully private LLM chatbot supporting Llama 3, Mistral and other open source models. \n\n- Fully private = No conversation data ever leaves your computer\n- Runs in the browser = No server needed and no install needed!\n- Works offline\n- Easy-to-use interface on par with ChatGPT, but for open source LLMs\n\nBig thanks to the inference engine provided by [webllm](https:\u002F\u002Fgithub.com\u002Fmlc-ai\u002Fweb-llm).\n\nJoin us on Discord\n\n[![](https:\u002F\u002Fdcbadge.vercel.app\u002Fapi\u002Fserver\u002FQkVzykMc9V)](https:\u002F\u002Fdiscord.gg\u002FQkVzykMc9V)\n\nhttps:\u002F\u002Fdiscord.gg\u002FQkVzykMc9V\n\n## System Requirements\n\nTo run this, you need a modern browser with support for WebGPU. According to [caniuse](https:\u002F\u002Fcaniuse.com\u002F?search=WebGPU), WebGPU is supported on:\n\n- Google Chrome\n- Microsoft Edge\n\nIt's also available in Firefox, but it needs to be enabled manually through the dom.webgpu.enabled flag. Safari on MacOS also has experimental support for WebGPU which can be enabled through the WebGPU experimental feature. \n\nIn addition to WebGPU support, various models might have specific RAM requirements.\n\n## Try it out\n\nYou can [try it here](https:\u002F\u002Fsecretllama.com).\n\nTo compile the React code yourself, download the repo and then, run\n\n```\nyarn\nyarn build-and-preview\n```\n\nIf you're looking to make changes, run the development environment with live reload:\n\n```\nyarn\nyarn dev\n```\n\n## Supported models\n\n| Model                     | Model Size |\n|---------------------------|------------|\n| TinyLlama-1.1B-Chat-v0.4-q4f32_1-1k | 600MB      |\n| Llama-3-8B-Instruct-q4f16_1 ⭐         | 4.3GB      |\n| Phi1.5-q4f16_1-1k                   | 1.2GB        |\n| Mistral-7B-Instruct-v0.2-q4f16_1 ⭐    | 4GB        |\n\n\n## Looking for contributors\n\nWe would love contributions to improve the interface, support more models, speed up initial model loading time and fix bugs.\n\n## Other Projects by Author\n\nCheck out [screenshot to code](https:\u002F\u002Fgithub.com\u002Fabi\u002Fscreenshot-to-code) and [Pico - AI-powered app builder](https:\u002F\u002Fpicoapps.xyz)\n","Secret Llama 是一个完全在浏览器中运行的私密大语言模型聊天机器人，无需服务器支持，并兼容Mistral和Llama 3等开源模型。其核心功能包括确保所有对话数据不离开用户设备、离线工作能力以及与ChatGPT相媲美的易用界面。该项目使用TypeScript编写，依赖WebGPU技术实现高性能计算，适用于需要高度隐私保护且偏好使用开源语言模型进行交互的应用场景，如个人研究、教育或对数据安全有严格要求的企业内部沟通工具。",2,"2026-06-11 03:47:31","high_star"]