{"id":2160,"date":"2026-08-10T20:23:12","date_gmt":"2026-08-10T20:23:12","guid":{"rendered":"https:\/\/convly.ai\/?p=2160"},"modified":"2026-08-10T20:23:12","modified_gmt":"2026-08-10T20:23:12","slug":"jan-ai-explained","status":"publish","type":"post","link":"https:\/\/convly.ai\/de\/jan-ai-explained\/","title":{"rendered":"Jan AI: Open-Source-Desktopanwendung zum lokalen Ausf\u00fchren von LLMs"},"content":{"rendered":"<div class=\"convly-tldr\">\n<ul>\n<li>Jan is a free, open-source desktop app (AGPL license) that runs LLMs entirely on your own hardware \u2014 no account, no cloud, no data leaving your machine.<\/li>\n<li>It ships a chat interface, a model hub for downloading GGUF models, and an OpenAI-compatible local API server (default port 1337).<\/li>\n<li>Download from <strong>jan.ai<\/strong> or the <a href=\"https:\/\/github.com\/janhq\/jan\/releases\" target=\"_blank\" rel=\"noopener\">GitHub Releases page<\/a> \u2014 builds for Windows, macOS (Apple Silicon and Intel), and Linux.<\/li>\n<li>Best for privacy-focused users who want a full GUI experience; developers wanting a headless API-first workflow may prefer Ollama instead.<\/li>\n<\/ul>\n<\/div>\n<p>Jan is an open-source desktop application for running large language models locally, built by Homebrew Research (janhq). It packages a chat interface, a model hub, and an inference engine into a single installable app. Everything runs offline: no account is required, no data is sent to external servers, and after your models are downloaded, no internet connection is needed at all. The source code lives at <strong>github.com\/janhq\/jan<\/strong> under the AGPL-3.0 license.<\/p>\n<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_86 counter-flat ez-toc-counter ez-toc-container-direction\">\n<label for=\"ez-toc-cssicon-toggle-item-6a7a5de0f3341\" class=\"ez-toc-cssicon-toggle-label\"><span class=\"\"><span class=\"eztoc-hide\" style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #000000;color:#000000\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewBox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #000000;color:#000000\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewBox=\"0 0 24 24\" version=\"1.2\" baseProfile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/label><input type=\"checkbox\"  id=\"ez-toc-cssicon-toggle-item-6a7a5de0f3341\"  aria-label=\"Toggle\" \/><nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/convly.ai\/de\/jan-ai-explained\/#Hardware_Requirements\" >Hardware Requirements<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/convly.ai\/de\/jan-ai-explained\/#Installing_Jan\" >Installing Jan<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/convly.ai\/de\/jan-ai-explained\/#Downloading_Your_First_Model\" >Downloading Your First Model<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/convly.ai\/de\/jan-ai-explained\/#The_Local_API_Server\" >The Local API Server<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/convly.ai\/de\/jan-ai-explained\/#Jan_vs_LM_Studio_vs_Ollama\" >Jan vs LM Studio vs Ollama<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/convly.ai\/de\/jan-ai-explained\/#Privacy_and_Offline_Use\" >Privacy and Offline Use<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-7\" href=\"https:\/\/convly.ai\/de\/jan-ai-explained\/#Frequently_Asked_Questions\" >Frequently Asked Questions<\/a><\/li><\/ul><\/nav><\/div>\n<h2><span class=\"ez-toc-section\" id=\"Hardware_Requirements\"><\/span>Hardware Requirements<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Jan can run on CPU alone, but a GPU makes a significant difference in response speed. The limiting resource is VRAM \u2014 how much GPU memory your card has determines which models you can load at usable speeds.<\/p>\n<table>\n<thead>\n<tr>\n<th>Component<\/th>\n<th>Minimum<\/th>\n<th>Recommended<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>System RAM<\/td>\n<td>8 GB<\/td>\n<td>16 GB+<\/td>\n<\/tr>\n<tr>\n<td>GPU VRAM<\/td>\n<td>None required<\/td>\n<td>8 GB+ for 7B models<\/td>\n<\/tr>\n<tr>\n<td>Storage<\/td>\n<td>10 GB free<\/td>\n<td>50 GB+ (models are large files)<\/td>\n<\/tr>\n<tr>\n<td>OS<\/td>\n<td>Windows 10, macOS 12, Ubuntu 20.04<\/td>\n<td>Latest stable versions<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>A 7B parameter model in 4-bit quantization (Q4_K_M) requires roughly 4\u20135 GB of VRAM; a 13B model needs around 8\u20139 GB. Use the <a href=\"https:\/\/convly.ai\/llm-vram-calculator\/\">Convly VRAM calculator<\/a> to estimate memory requirements for your specific model and quantization before downloading. If you are choosing hardware, the <a href=\"https:\/\/convly.ai\/best-gpus-for-local-llms-2026\/\">best GPUs for local LLMs guide<\/a> covers current options across price tiers.<\/p>\n<p>On Apple Silicon Macs, Jan activates Metal acceleration automatically. On Windows and Linux with an NVIDIA card, it uses CUDA. AMD GPU support is available via Vulkan on Linux.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Installing_Jan\"><\/span>Installing Jan<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Download the installer from <strong>jan.ai<\/strong> or the <a href=\"https:\/\/github.com\/janhq\/jan\/releases\" target=\"_blank\" rel=\"noopener\">GitHub Releases page<\/a>. Jan bundles its own llama.cpp inference engine \u2014 no separate dependencies required.<\/p>\n<h3>Windows<\/h3>\n<p>Run the <code>.exe<\/code> installer. Jan installs to <code>%LOCALAPPDATA%ProgramsJan<\/code> by default. No administrator rights required.<\/p>\n<h3>macOS<\/h3>\n<p>Download the <code>.dmg<\/code>. Separate builds exist for Apple Silicon and Intel \u2014 pick the one matching your Mac. Drag Jan to <code>\/Applications<\/code>. On first launch, right-click the icon and choose <strong>Open<\/strong> to bypass the Gatekeeper warning.<\/p>\n<h3>Linux<\/h3>\n<p>Jan provides an <code>.AppImage<\/code>, a <code>.deb<\/code>, and an <code>.rpm<\/code> on the releases page.<\/p>\n<pre><code># AppImage\nchmod +x jan-linux-x86_64-*.AppImage\n.\/jan-linux-x86_64-*.AppImage\n\n# Debian\/Ubuntu\nsudo dpkg -i jan-linux-amd64-*.deb<\/code><\/pre>\n<h2><span class=\"ez-toc-section\" id=\"Downloading_Your_First_Model\"><\/span>Downloading Your First Model<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>When Jan opens, the <strong>Hub<\/strong> tab shows a curated browser of GGUF models from Hugging Face: Llama, Mistral, Qwen, Gemma, Phi, and others in multiple quantizations.<\/p>\n<ol>\n<li>Click <strong>Hub<\/strong> in the left sidebar.<\/li>\n<li>Search for a model \u2014 &#8220;Llama 3.2&#8221; is a good starting point for most hardware.<\/li>\n<li>Select a quantization. <strong>Q4_K_M<\/strong> is the default recommendation: smaller than Q6\/Q8 with acceptable quality loss.<\/li>\n<li>Click <strong>Download<\/strong> and wait for the file transfer to complete.<\/li>\n<\/ol>\n<p>Jan stores downloaded models at <code>~\/jan\/models\/<\/code> on macOS and Linux, and at <code>C:Users&lt;username&gt;janmodels<\/code> on Windows. Each model lives in its own named subfolder containing the <code>.gguf<\/code> file and a <code>model.json<\/code> metadata file. You can also drop any third-party <code>.gguf<\/code> file into a new folder there \u2014 Jan detects it on next launch. GGUF files from LM Studio are directly compatible.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"The_Local_API_Server\"><\/span>The Local API Server<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Jan includes an OpenAI-compatible REST API server, letting any tool that supports the OpenAI client library \u2014 the Python <code>openai<\/code> SDK, LangChain, Continue.dev, and others \u2014 use Jan as a local backend.<\/p>\n<p>To start it: open <strong>Settings &rarr; Local API Server<\/strong>, toggle the server on. It listens on <strong>port 1337<\/strong> by default; this is configurable in that same panel.<\/p>\n<pre><code>from openai import OpenAI\n\nclient = OpenAI(\n    base_url=\"http:\/\/localhost:1337\/v1\",\n    api_key=\"jan\"  # any non-empty string\n)\n\nresponse = client.chat.completions.create(\n    model=\"llama3.2-3b-instruct\",  # matches folder name under ~\/jan\/models\/\n    messages=[{\"role\": \"user\", \"content\": \"Hello\"}]\n)\nprint(response.choices[0].message.content)<\/code><\/pre>\n<p>The model name in API calls must match the folder name Jan uses internally \u2014 visible in the Hub or in the model dropdown in Chat view.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Jan_vs_LM_Studio_vs_Ollama\"><\/span>Jan vs LM Studio vs Ollama<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>All three tools run GGUF models locally. The differences are in philosophy and workflow.<\/p>\n<table>\n<thead>\n<tr>\n<th><\/th>\n<th>Jan<\/th>\n<th>LM Studio<\/th>\n<th>Ollama<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Interface<\/td>\n<td>Full GUI, chat-first<\/td>\n<td>Full GUI, chat-first<\/td>\n<td>CLI only (third-party GUIs available)<\/td>\n<\/tr>\n<tr>\n<td>License<\/td>\n<td>AGPL-3.0 (open source)<\/td>\n<td>Proprietary freeware<\/td>\n<td>MIT (open source)<\/td>\n<\/tr>\n<tr>\n<td>Local API<\/td>\n<td>Yes, OpenAI-compatible (port 1337)<\/td>\n<td>Yes, OpenAI-compatible (port 1234)<\/td>\n<td>Yes, own API + OpenAI-compatible endpoint<\/td>\n<\/tr>\n<tr>\n<td>Model management<\/td>\n<td>Built-in Hub (Hugging Face)<\/td>\n<td>Built-in search (Hugging Face)<\/td>\n<td>CLI pull from Ollama registry<\/td>\n<\/tr>\n<tr>\n<td>Model format<\/td>\n<td>GGUF<\/td>\n<td>GGUF<\/td>\n<td>Ollama format (wraps GGUF internally)<\/td>\n<\/tr>\n<tr>\n<td>Headless\/server use<\/td>\n<td>Limited<\/td>\n<td>Limited<\/td>\n<td>Excellent \u2014 designed for it<\/td>\n<\/tr>\n<tr>\n<td>System prompt \/ persona<\/td>\n<td>Yes, per-assistant<\/td>\n<td>Yes, per-model preset<\/td>\n<td>Via Modelfile<\/td>\n<\/tr>\n<tr>\n<td>Platform<\/td>\n<td>Windows, macOS, Linux<\/td>\n<td>Windows, macOS, Linux<\/td>\n<td>Windows, macOS, Linux<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><strong>Choose Jan<\/strong> if you want a fully open-source GUI tool and care about the AGPL guarantee that no one is building a closed product on top of the code you use. Jan&#8217;s chat interface is clean and the extension system lets you add capabilities without touching config files.<\/p>\n<p><strong>Choose LM Studio<\/strong> if polish matters more than open-source licensing. LM Studio has a more refined UI and slightly broader hardware compatibility documentation. See the <a href=\"https:\/\/convly.ai\/lm-studio-complete-guide-2026\/\">LM Studio complete guide<\/a> for a full walkthrough.<\/p>\n<p><strong>Choose Ollama<\/strong> if you are a developer who wants a fast, scriptable backend with no GUI overhead. Ollama is easier to integrate into CI pipelines, shell scripts, or Docker containers. The <a href=\"https:\/\/convly.ai\/what-is-ollama-complete-guide-2026\/\">Ollama complete guide<\/a> covers that path in detail.<\/p>\n<p>If you are comparing the running costs of local inference against using a cloud API, the <a href=\"https:\/\/convly.ai\/self-hosting-vs-api-calculator\/\">self-hosting vs API break-even calculator<\/a> helps quantify the trade-off based on your actual usage volume and hardware costs.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Privacy_and_Offline_Use\"><\/span>Privacy and Offline Use<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Jan&#8217;s core privacy guarantee is architectural: inference happens inside llama.cpp on your local machine. There is no telemetry sent during inference, no model query logged to a remote server, and no account required. Once a model file is downloaded, you can run Jan with no network connection indefinitely.<\/p>\n<p>This makes Jan suitable for working with documents that cannot leave your machine \u2014 legal, medical, or commercially sensitive text \u2014 in a way that cloud API tools cannot match regardless of their privacy policies.<\/p>\n<p>Jan does check for app updates and fetch the Hub model list from the internet when a connection is available. Both can be disabled: go to <strong>Settings &rarr; Advanced<\/strong> to turn off update checks, and the Hub simply shows no new models when offline without breaking the rest of the app.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Frequently_Asked_Questions\"><\/span>Frequently Asked Questions<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<h3>Is Jan AI completely free?<\/h3>\n<p>Yes. Jan is free to download and use with no subscription, no usage limits, and no paywalled features. The source code is published under AGPL-3.0 on GitHub. The only cost is your own hardware and electricity.<\/p>\n<h3>What models can Jan run?<\/h3>\n<p>Jan runs any model in GGUF format. This includes most popular open-weight models: the Llama family, Mistral, Mixtral, Qwen, Gemma, Phi, DeepSeek, and many others. The built-in Hub surfaces a curated subset; you can add any GGUF file manually by placing it in the models directory.<\/p>\n<h3>How much VRAM do I need?<\/h3>\n<p>It depends on model size and quantization. A 7B model at Q4_K_M needs roughly 4\u20135 GB of VRAM; a 13B model needs 8\u20139 GB. CPU-only inference works but is slow \u2014 expect 2\u20135 tokens per second on a modern CPU versus 30\u201380+ on a mid-range GPU. The <a href=\"https:\/\/convly.ai\/vram-requirements-every-major-llm-2026\/\">VRAM requirements guide<\/a> lists figures for major models.<\/p>\n<h3>Can I use Jan with VS Code or other coding tools?<\/h3>\n<p>Yes. Enable Jan&#8217;s local API server and point any OpenAI-compatible tool at <code>http:\/\/localhost:1337\/v1<\/code>. Continue.dev, Aider, and similar coding assistants all support custom base URLs. Set the model name to match the folder name Jan uses for your downloaded model.<\/p>\n<h3>What is the difference between Jan and Ollama?<\/h3>\n<p>Jan is a desktop GUI application \u2014 you interact through a chat interface and manage models through a visual hub. Ollama is a CLI tool and background service designed for developers who want to call models from scripts or integrate them into applications. Both expose an OpenAI-compatible API. Jan suits non-technical users and GUI-first workflows; Ollama suits automation and server deployments.<\/p>\n<h3>Does Jan support multi-modal models (images)?<\/h3>\n<p>Jan has added experimental support for vision models in recent releases, depending on the version installed. Check the Hub for models labeled with vision or multimodal capability. Support and model availability vary by release, so consult the GitHub release notes for your installed version rather than assuming a specific model will work.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Jan is a free, open-source desktop app (AGPL license) that runs LLMs entirely on your own hardware \u2014 no account, [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":2161,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[7],"tags":[],"class_list":["post-2160","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-news"],"_links":{"self":[{"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/posts\/2160","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/comments?post=2160"}],"version-history":[{"count":1,"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/posts\/2160\/revisions"}],"predecessor-version":[{"id":2162,"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/posts\/2160\/revisions\/2162"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/media\/2161"}],"wp:attachment":[{"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/media?parent=2160"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/categories?post=2160"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/convly.ai\/de\/wp-json\/wp\/v2\/tags?post=2160"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}