{"id":59,"date":"2026-05-18T12:37:28","date_gmt":"2026-05-18T12:37:28","guid":{"rendered":"https:\/\/convly.ai\/run-llama3-locally-laptop\/"},"modified":"2026-06-10T05:06:06","modified_gmt":"2026-06-10T05:06:06","slug":"run-llama3-locally-laptop","status":"publish","type":"post","link":"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/","title":{"rendered":"Come eseguire Llama in locale sul proprio laptop nel 2026 (guida completa alla configurazione)"},"content":{"rendered":"<p>Running a large language model on your own laptop used to be a research project. In 2026 it&#8217;s a 15-minute setup. You can have a genuinely capable AI assistant running entirely on your machine \u2014 no subscription, no internet required, and no data ever leaving your computer.<\/p>\n<p>This guide walks through the whole process: what hardware you need, which tool to use, which model to download, and how to get it running.<\/p>\n<div class=\"convly-tldr\">\n<h3>Punti chiave<\/h3>\n<ul>\n<li><strong>Easiest path:<\/strong> install Ollama or LM Studio \u2014 both get you running in minutes.<\/li>\n<li><strong>Hardware:<\/strong> 16 GB of RAM is the comfortable minimum; an Apple Silicon Mac or a laptop with a discrete GPU is ideal.<\/li>\n<li><strong>Model size:<\/strong> 7\u20138B models are the sweet spot for laptops \u2014 capable and fast.<\/li>\n<li><strong>Quantization<\/strong> shrinks models to fit your hardware; &#8220;Q4&#8221; versions are the standard choice.<\/li>\n<li><strong>Why do it:<\/strong> it&#8217;s free, fully private, and works offline.<\/li>\n<\/ul>\n<\/div>\n<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_84 counter-flat ez-toc-counter ez-toc-container-direction\">\n<label for=\"ez-toc-cssicon-toggle-item-6a38a8e480e2a\" class=\"ez-toc-cssicon-toggle-label\"><span class=\"\"><span class=\"eztoc-hide\" style=\"display:none;\">Attiva\/Disattiva<\/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-6a38a8e480e2a\"  aria-label=\"Attiva\/Disattiva\" \/><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\/it\/run-llama3-locally-laptop\/#Why_run_an_LLM_locally\" >Why run an LLM locally?<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#Step_1_Check_your_hardware\" >Step 1: Check your hardware<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#Step_2_Choose_your_tool\" >Step 2: Choose your tool<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#Step_3_Install_and_run_your_first_model\" >Step 3: Install and run your first model<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#Step_4_Pick_the_right_model_and_size\" >Step 4: Pick the right model and size<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#Step_5_Understand_quantization\" >Step 5: Understand quantization<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-7\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#Going_further\" >Going further<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-8\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#Why_its_slow_%E2%80%94_and_how_to_fix_it\" >Why it&#8217;s slow \u2014 and how to fix it<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-9\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#FAQ\" >Domande frequenti<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-10\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#Bottom_line\" >Conclusione<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-11\" href=\"https:\/\/convly.ai\/it\/run-llama3-locally-laptop\/#Related_articles\" >Articoli correlati<\/a><\/li><\/ul><\/nav><\/div>\n<h2><span class=\"ez-toc-section\" id=\"Why_run_an_LLM_locally\"><\/span>Why run an LLM locally?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Cloud AI is convenient, so why run a model yourself? Three real reasons:<\/p>\n<ul>\n<li><strong>Privacy.<\/strong> Nothing you type leaves your machine. For sensitive, confidential, or personal work, that&#8217;s a genuine advantage.<\/li>\n<li><strong>Cost.<\/strong> It&#8217;s free. No subscription, no per-token billing, no usage caps \u2014 generate as much as you like.<\/li>\n<li><strong>Offline and always available.<\/strong> It works on a plane, with no internet, and it can&#8217;t be rate-limited or discontinued.<\/li>\n<\/ul>\n<p>The trade-off: a model that runs on a laptop is smaller and less capable than a frontier cloud model. But modern small models are good enough for a lot of real work \u2014 writing, summarizing, coding help, brainstorming, Q&amp;A.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Step_1_Check_your_hardware\"><\/span>Step 1: Check your hardware<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Local LLM performance depends mostly on memory. Here&#8217;s the honest picture:<\/p>\n<table class=\"convly-vs\">\n<thead>\n<tr>\n<th>Your laptop<\/th>\n<th>What you can run<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>8 GB RAM<\/td>\n<td>Small models only (1\u20133B). Usable but limited.<\/td>\n<\/tr>\n<tr>\n<td>16 GB RAM<\/td>\n<td>7\u20138B models comfortably \u2014 the sweet spot.<\/td>\n<\/tr>\n<tr>\n<td>32 GB RAM<\/td>\n<td>Up to ~13\u201314B models with good speed.<\/td>\n<\/tr>\n<tr>\n<td>Apple Silicon (M-series)<\/td>\n<td>Excellent \u2014 unified memory is ideal; larger models run well.<\/td>\n<\/tr>\n<tr>\n<td>Discrete NVIDIA GPU<\/td>\n<td>Fastest option; VRAM is the limit for model size.<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>The two things that matter: <strong>total memory<\/strong> (RAM, or VRAM on a GPU) sets the largest model you can load, and a <strong>GPU or Apple Silicon<\/strong> sets how fast it runs. A modern laptop with 16 GB of RAM is a perfectly good starting point.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Step_2_Choose_your_tool\"><\/span>Step 2: Choose your tool<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>You don&#8217;t interact with the raw model \u2014 you use a tool that downloads, manages, and runs it. The best options in 2026:<\/p>\n<ul>\n<li><strong>Ollama<\/strong> \u2014 the most popular choice. A clean command-line tool (with a simple app) that downloads and runs models with a single command, and exposes a local API so other apps can connect. Best all-round pick.<\/li>\n<li><strong>LM Studio<\/strong> \u2014 a polished graphical app. Browse and download models, chat in a built-in interface, no command line needed. Best for beginners who want a visual experience.<\/li>\n<li><strong>Gen<\/strong> \u2014 an open-source, privacy-focused desktop app, a clean alternative to LM Studio.<\/li>\n<li><strong>llama.cpp<\/strong> \u2014 the high-performance engine many of these tools are built on. Use it directly if you want maximum control and efficiency.<\/li>\n<\/ul>\n<p>For most people: <strong>Ollama<\/strong> if you&#8217;re comfortable with a terminal, <strong>LM Studio<\/strong> if you&#8217;d rather click.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Step_3_Install_and_run_your_first_model\"><\/span>Step 3: Install and run your first model<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The setup with Ollama is genuinely this short:<\/p>\n<ol>\n<li>Download and install Ollama from its official site.<\/li>\n<li>Open a terminal.<\/li>\n<li>Run one command:<\/li>\n<\/ol>\n<pre><code>ollama run llama3.1\n<\/code><\/pre>\n<p>That command downloads the model the first time (a few gigabytes) and then drops you into a chat prompt. That&#8217;s it \u2014 you now have a private AI assistant running locally. The next time, it starts instantly.<\/p>\n<p>With LM Studio the equivalent is: open the app, search for a model, click download, then click to start chatting \u2014 entirely through the interface.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Step_4_Pick_the_right_model_and_size\"><\/span>Step 4: Pick the right model and size<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Two things to choose: the model family and its size.<\/p>\n<p><strong>Model family<\/strong> \u2014 strong open models that run well locally include Meta&#8217;s <strong>Llama<\/strong> series, Alibaba&#8217;s <strong>Qwen<\/strong>, Google&#8217;s <strong>Gemma<\/strong>, Mistral&#8217;s models, and DeepSeek&#8217;s smaller releases. They&#8217;re all good; try a couple and see which you prefer.<\/p>\n<p><strong>Size<\/strong> \u2014 models come in parameter counts marked like 3B, 8B, 14B (B = billion):<\/p>\n<ul>\n<li><strong>1\u20133B<\/strong> \u2014 very fast, light on memory, fine for simple tasks. Good for 8 GB machines.<\/li>\n<li><strong>7\u20138B<\/strong> \u2014 the laptop sweet spot. Genuinely capable for writing, coding help, and Q&amp;A, and runs well on 16 GB.<\/li>\n<li><strong>13\u201314B and up<\/strong> \u2014 noticeably smarter, but need 32 GB or a strong GPU.<\/li>\n<\/ul>\n<p>Start with an 8B model. It&#8217;s the best balance of capability and speed for most laptops.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Step_5_Understand_quantization\"><\/span>Step 5: Understand quantization<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>You&#8217;ll see model names with tags like <code>Q4_K_M<\/code> or <code>Q8<\/code>. This is <strong>quantizzazione<\/strong> \u2014 a compression technique that reduces the precision of the model&#8217;s numbers so it uses far less memory, with only a small quality loss.<\/p>\n<ul>\n<li><strong>Q8<\/strong> \u2014 highest quality, largest size.<\/li>\n<li><strong>Q4<\/strong> \u2014 about half the memory of Q8, with quality that&#8217;s very close. <strong>This is the standard recommendation.<\/strong><\/li>\n<li><strong>Q2\/Q3<\/strong> \u2014 smallest, but quality degrades noticeably; use only if memory forces it.<\/li>\n<\/ul>\n<p>The practical rule: choose a <strong>Q4<\/strong> version of the largest model your memory can comfortably hold. Tools like Ollama pick a sensible quantization by default, so you often don&#8217;t have to think about it.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Going_further\"><\/span>Going further<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Once it&#8217;s running, you can do more than chat in a terminal:<\/p>\n<ul>\n<li><strong>Connect a nicer interface<\/strong> \u2014 apps like Open WebUI give a ChatGPT-style window over your local model.<\/li>\n<li><strong>Use the local API<\/strong> \u2014 Ollama serves an API on your machine, so you can build scripts and apps against your local model exactly as you would a cloud one.<\/li>\n<li><strong>Try retrieval<\/strong> \u2014 point a <a href=\"\/it\/rag-retrieval-augmented-generation-explained\/\">RAG setup<\/a> at your own documents for a fully private &#8220;chat with your files&#8221; assistant.<\/li>\n<\/ul>\n<p><!--ai-enriched--><\/p>\n<h2><span class=\"ez-toc-section\" id=\"Why_its_slow_%E2%80%94_and_how_to_fix_it\"><\/span>Why it&#8217;s slow \u2014 and how to fix it<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The most common complaint after a first install isn&#8217;t that the model won&#8217;t run \u2014 it&#8217;s that it crawls. On a laptop, slow output almost always comes down to the model not actually using your GPU. The fastest way to check is to run a model, then in another terminal run <strong>ollama ps<\/strong>. The output shows how the model is split: if it reports 100% GPU you&#8217;re fine; if it shows 100% CPU or a CPU\/GPU split, you&#8217;ve found your problem.<\/p>\n<p>There are three usual culprits, in order of how often they bite:<\/p>\n<ul>\n<li><strong>The GPU was never detected.<\/strong> On Windows and Linux with an NVIDIA card, this usually means the GPU drivers were installed <em>after<\/em> the runtime, so it never picked up CUDA support \u2014 Ollama checks for the GPU at install time, not while it runs. Confirm <strong>nvidia-smi<\/strong> works, then reinstall the runtime so it detects the GPU. This single fix resolves the majority of &#8220;it&#8217;s using my CPU&#8221; reports.<\/li>\n<li><strong>The model is too big for your VRAM.<\/strong> When a model doesn&#8217;t fit, the overflow layers silently fall back to system RAM and CPU \u2014 and those few CPU layers drag the whole thing down. The fix is to drop to a smaller model or a heavier quantization (a lower-Q build) so the full model lives in VRAM.<\/li>\n<li><strong>Your context window is too large.<\/strong> A long context consumes memory too, because the KV cache grows with it. Push it too high and layers spill back onto the CPU. If you don&#8217;t need a huge prompt, lower the context length (8K is plenty for most work) and the model fits more comfortably.<\/li>\n<\/ul>\n<p>Two problems are specific to laptops. First, <strong>battery power policy<\/strong>: most Windows laptops throttle the discrete GPU hard when unplugged, and unplugging can cut inference speed by half or more. This is firmware behavior, not a bug \u2014 keep the laptop plugged in for serious work. Second, <strong>thermal throttling<\/strong>: after roughly 10\u201320 minutes of sustained generation, a thin laptop heats up and clocks down. Raising the laptop a couple of centimetres on a stand for airflow, and preferring a lighter quantization that runs cooler, both push back the point where throttling kicks in. None of this makes a laptop a workstation, but it&#8217;s the difference between a few tokens per second and a genuinely usable assistant.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"FAQ\"><\/span>Domande frequenti<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<h3>Can I run Llama on a normal laptop?<\/h3>\n<p>Yes. A laptop with 16 GB of RAM comfortably runs 7\u20138B models, which are genuinely useful. Even 8 GB machines can run smaller 1\u20133B models. Apple Silicon Macs and laptops with a discrete GPU run local models especially well.<\/p>\n<h3>Is running an LLM locally free?<\/h3>\n<p>Yes. The models are free to download and there&#8217;s no usage cost \u2014 you can generate as much as you want. The only &#8220;cost&#8221; is your hardware and the disk space the model files take up (a few gigabytes each).<\/p>\n<h3>What is the best tool to run LLMs locally?<\/h3>\n<p>Ollama is the most popular and the best all-round choice \u2014 a simple command downloads and runs any model, and it provides a local API. LM Studio is the best option if you prefer a graphical app with no command line.<\/p>\n<h3>How much RAM do I need to run a local LLM?<\/h3>\n<p>16 GB is the comfortable minimum for genuinely capable 7\u20138B models. With 8 GB you&#8217;re limited to smaller 1\u20133B models. With 32 GB you can run 13\u201314B models. More memory mostly lets you run larger, smarter models.<\/p>\n<h3>Are local LLMs as good as ChatGPT?<\/h3>\n<p>Not as capable as a frontier cloud model \u2014 laptop-sized models are smaller and less powerful. But they are good enough for many everyday tasks: writing, summarizing, coding assistance, and Q&amp;A. You trade some capability for total privacy, zero cost, and offline access.<\/p>\n<h3>Why is my local LLM so slow?<\/h3>\n<p>Nine times out of ten, the model isn&#8217;t using your GPU. Run <strong>ollama ps<\/strong> while a model is loaded: if it shows 100% CPU or a CPU\/GPU split, that&#8217;s your answer. The usual causes are GPU drivers installed after the runtime (reinstall the runtime so it picks up CUDA), a model too large for your VRAM (use a smaller model or heavier quantization), or a context window so big it pushes layers onto the CPU (lower it).<\/p>\n<h3>Should I keep my laptop plugged in while running a local LLM?<\/h3>\n<p>Yes, for anything beyond a quick question. Most Windows laptops aggressively limit the discrete GPU on battery to preserve runtime, which can roughly halve your tokens-per-second. That slowdown is firmware power policy, not a fault. Plugging in restores full GPU clocks; a cooling stand for airflow helps avoid the thermal throttling that creeps in after extended sessions.<\/p>\n<h3>Can I use a local LLM completely offline?<\/h3>\n<p>Yes. Only the initial model download needs internet. Once the model is on disk, it runs fully offline \u2014 you can disconnect entirely and it still responds. That&#8217;s the core privacy benefit: your prompts never leave the machine, which makes a local model a sensible choice for confidential notes, draft work, or anything you wouldn&#8217;t want sent to a cloud service.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Bottom_line\"><\/span>Conclusione<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Running an AI model on your own laptop is no longer difficult. Install <strong>Ollama<\/strong> or <strong>LM Studio<\/strong>, download an <strong>8B model<\/strong> in a <strong>Q4<\/strong> quantization, and within 15 minutes you have a capable assistant that&#8217;s free, fully private, and works offline.<\/p>\n<p>It won&#8217;t replace a frontier cloud model for the hardest tasks \u2014 but for everyday writing, coding help, and private Q&amp;A, a local model is genuinely useful. And once it&#8217;s running, you own it: no subscription, no limits, and no data leaving your machine.<\/p>\n<p><!--related-block--><\/p>\n<div class=\"convly-related\">\n<h2><span class=\"ez-toc-section\" id=\"Related_articles\"><\/span>Articoli correlati<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<ul>\n<li><a href=\"https:\/\/convly.ai\/it\/90-day-ai-engineer-path\/\">Da zero a ingegnere AI: il tuo percorso di apprendimento di 90 giorni<\/a><\/li>\n<li><a href=\"https:\/\/convly.ai\/it\/local-llm-ollama-setup\/\">Configurare il primo modello linguistico locale con Ollama<\/a><\/li>\n<li><a href=\"https:\/\/convly.ai\/it\/build-personal-ai-assistant-python\/\">Build a Personal AI Assistant in 30 Minutes (Python Tutorial)<\/a><\/li>\n<li><a href=\"https:\/\/convly.ai\/it\/how-to-run-llama-3-locally-on-snapdragon-8-gen-4\/\">How to Run Llama 3 Locally on Snapdragon 8 Gen 4 (Step-by-Step, 2026)<\/a><\/li>\n<\/ul>\n<\/div>","protected":false},"excerpt":{"rendered":"<p>Run a capable AI model on your own laptop \u2014 free, private, and offline. This step-by-step guide covers the hardware you need, the best tools, and which model size to pick.<\/p>","protected":false},"author":0,"featured_media":60,"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":[9],"tags":[260,256,458,259,457],"class_list":["post-59","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-tutorials","tag-lm-studio","tag-local-llm","tag-offline-ai","tag-ollama","tag-run-llama-locally"],"_links":{"self":[{"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/posts\/59","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/comments?post=59"}],"version-history":[{"count":3,"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/posts\/59\/revisions"}],"predecessor-version":[{"id":1042,"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/posts\/59\/revisions\/1042"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/media\/60"}],"wp:attachment":[{"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/media?parent=59"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/categories?post=59"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/convly.ai\/it\/wp-json\/wp\/v2\/tags?post=59"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}