{"id":2725,"date":"2026-09-16T20:13:30","date_gmt":"2026-09-16T20:13:30","guid":{"rendered":"https:\/\/convly.ai\/?p=2725"},"modified":"2026-09-16T20:13:30","modified_gmt":"2026-09-16T20:13:30","slug":"hugging-face-image-to-video","status":"publish","type":"post","link":"https:\/\/convly.ai\/fr\/hugging-face-image-to-video\/","title":{"rendered":"Hugging Face Image to Video : mod\u00e8les, Spaces et inf\u00e9rence locale"},"content":{"rendered":"<div class=\"convly-tldr\">\n<ul>\n<li>Hugging Face hosts image-to-video (I2V) models you can run three ways: hosted <strong>Spaces<\/strong> demos, the <strong>Inference API\/Endpoints<\/strong>, or locally with the <code>diffusers<\/code> library.<\/li>\n<li>Popular open-weight I2V checkpoints include <strong>Wan 2.1\/2.2 I2V<\/strong> (Alibaba), <strong>Stable Video Diffusion (SVD\/SVD-XT)<\/strong> from Stability AI, and <strong>LTX-Video<\/strong> from Lightricks.<\/li>\n<li>Realistic local VRAM: ~10\u201316 GB for SVD at 576\u00d71024, 24 GB+ for Wan 2.2 I2V-A14B at 720p, with offloading tricks in <code>diffusers<\/code> for smaller cards.<\/li>\n<li>If you don&#8217;t want to manage GPUs, hosted APIs like <a href=\"https:\/\/convly.ai\/model\/wan-2-5\/\">Wan 2.5<\/a> ($0.05\/sec) or <a href=\"https:\/\/convly.ai\/model\/kling-2-5-turbo-pro\/\">Kling 2.5 Turbo Pro<\/a> ($0.07\/sec) are usually cheaper than buying a GPU.<\/li>\n<\/ul>\n<\/div>\n<p><strong>Hugging Face image to video<\/strong> refers to using models hosted on <a href=\"https:\/\/huggingface.co\/models?pipeline_tag=image-to-video\" rel=\"noopener\" target=\"_blank\">huggingface.co under the <code>image-to-video<\/code> pipeline tag<\/a> to turn a still image into a short animated clip. You interact with them in three ways: click-through <strong>Spaces<\/strong> demos in the browser, the <strong>Inference API<\/strong>\/Inference Endpoints, or by downloading the weights and running them locally with the <code>diffusers<\/code> Python library. Below is what actually works, per model, per platform.<\/p>\n<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_87_1 counter-flat ez-toc-counter ez-toc-container-direction\">\n<label for=\"ez-toc-cssicon-toggle-item-6aab258192b1b\" 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-6aab258192b1b\"  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\/fr\/hugging-face-image-to-video\/#What_the_image-to-video_tag_actually_covers\" >What the image-to-video tag actually covers<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/convly.ai\/fr\/hugging-face-image-to-video\/#Three_ways_to_run_Hugging_Face_image_to_video\" >Three ways to run Hugging Face image to video<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/convly.ai\/fr\/hugging-face-image-to-video\/#VRAM_and_hardware_honestly\" >VRAM and hardware, honestly<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/convly.ai\/fr\/hugging-face-image-to-video\/#Windows\" >Windows<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/convly.ai\/fr\/hugging-face-image-to-video\/#macOS\" >macOS<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/convly.ai\/fr\/hugging-face-image-to-video\/#Linux\" >Linux<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-7\" href=\"https:\/\/convly.ai\/fr\/hugging-face-image-to-video\/#When_to_skip_Hugging_Face_and_use_a_hosted_video_API\" >When to skip Hugging Face and use a hosted video API<\/a><\/li><li class='ez-toc-page-1'><a class=\"ez-toc-link ez-toc-heading-8\" href=\"https:\/\/convly.ai\/fr\/hugging-face-image-to-video\/#Frequently_asked_questions\" >Frequently asked questions<\/a><\/li><\/ul><\/nav><\/div>\n<h2><span class=\"ez-toc-section\" id=\"What_the_image-to-video_tag_actually_covers\"><\/span>What the <code>image-to-video<\/code> tag actually covers<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The pipeline tag filters the Hub to models whose intended input is one image (optionally plus a text prompt) and whose output is a short video, typically 2\u20136 seconds at 16\u201324 fps. As of writing, the most active open-weight families are:<\/p>\n<table>\n<thead>\n<tr>\n<th>Model<\/th>\n<th>Author<\/th>\n<th>Typical output<\/th>\n<th>Weights license<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Wan 2.2 I2V-A14B<\/td>\n<td>Alibaba (Wan-AI)<\/td>\n<td>720p, 5s, 24 fps<\/td>\n<td>Apache 2.0<\/td>\n<\/tr>\n<tr>\n<td>Wan 2.1 I2V-14B \/ I2V-1.3B<\/td>\n<td>Alibaba (Wan-AI)<\/td>\n<td>480p\u2013720p, ~5s<\/td>\n<td>Apache 2.0<\/td>\n<\/tr>\n<tr>\n<td>Stable Video Diffusion (SVD)<\/td>\n<td>Stability AI<\/td>\n<td>576\u00d71024, 14 frames<\/td>\n<td>SVD non-commercial \/ commercial via Stability<\/td>\n<\/tr>\n<tr>\n<td>SVD-XT<\/td>\n<td>Stability AI<\/td>\n<td>576\u00d71024, 25 frames<\/td>\n<td>Same as SVD<\/td>\n<\/tr>\n<tr>\n<td>LTX-Video<\/td>\n<td>Lightricks<\/td>\n<td>768\u00d7512, ~5s at 24 fps<\/td>\n<td>OpenRAIL-M variant<\/td>\n<\/tr>\n<tr>\n<td>CogVideoX-5B-I2V<\/td>\n<td>THUDM<\/td>\n<td>720\u00d7480, 6s at 8 fps<\/td>\n<td>Apache 2.0 (code), custom (weights)<\/td>\n<\/tr>\n<tr>\n<td>image-to-video<\/td>\n<td>Bantikumar (community)<\/td>\n<td>Small demo checkpoint<\/td>\n<td>See model card<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>Always confirm resolution, frame count and license on the model card itself &mdash; authors update these. The Wan 2.2 card is at <a href=\"https:\/\/huggingface.co\/Wan-AI\/Wan2.2-I2V-A14B\" rel=\"noopener\" target=\"_blank\">huggingface.co\/Wan-AI\/Wan2.2-I2V-A14B<\/a>, SVD-XT at <a href=\"https:\/\/huggingface.co\/stabilityai\/stable-video-diffusion-img2vid-xt\" rel=\"noopener\" target=\"_blank\">huggingface.co\/stabilityai\/stable-video-diffusion-img2vid-xt<\/a>, and LTX-Video at <a href=\"https:\/\/huggingface.co\/Lightricks\/LTX-Video\" rel=\"noopener\" target=\"_blank\">huggingface.co\/Lightricks\/LTX-Video<\/a>.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Three_ways_to_run_Hugging_Face_image_to_video\"><\/span>Three ways to run Hugging Face image to video<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<h3>1. Spaces (no install, browser only)<\/h3>\n<p>Every major I2V model has one or more community Spaces. Go to <a href=\"https:\/\/huggingface.co\/spaces\" rel=\"noopener\" target=\"_blank\">huggingface.co\/spaces<\/a>, search for the model name (e.g. &#8220;Wan 2.2 I2V&#8221; or &#8220;LTX-Video&#8221;), upload an image, type an optional prompt, click <em>Generate<\/em>. Free Spaces run on shared ZeroGPU (H200 slices) with a per-user daily quota; if the Space is busy you&#8217;ll queue. For production, duplicate the Space and attach paid hardware from the Space settings.<\/p>\n<h3>2. Inference API and Inference Endpoints<\/h3>\n<p>The serverless <a href=\"https:\/\/huggingface.co\/docs\/api-inference\/index\" rel=\"noopener\" target=\"_blank\">Inference API<\/a> supports some I2V models directly; for others you deploy a dedicated <a href=\"https:\/\/huggingface.co\/docs\/inference-endpoints\/index\" rel=\"noopener\" target=\"_blank\">Inference Endpoint<\/a> on an A10G, L4, L40S, A100 or H100. Endpoints are billed per hour of GPU uptime, so leaving one running for a month is expensive; a 24\/7 A100-80GB endpoint typically runs into thousands of dollars per month. Use the <a href=\"https:\/\/convly.ai\/self-hosting-vs-api-calculator\/\">self-hosting vs API calculator<\/a> to compare against per-second video APIs.<\/p>\n<h3>3. Local inference with <code>diffusers<\/code><\/h3>\n<p>The reference client is the Hugging Face <a href=\"https:\/\/huggingface.co\/docs\/diffusers\/index\" rel=\"noopener\" target=\"_blank\">diffusers<\/a> library. Install:<\/p>\n<pre><code>pip install --upgrade diffusers transformers accelerate torch imageio imageio-ffmpeg<\/code><\/pre>\n<p>Minimal SVD-XT example (from the model card):<\/p>\n<pre><code>import torch\nfrom diffusers import StableVideoDiffusionPipeline\nfrom diffusers.utils import load_image, export_to_video\n\npipe = StableVideoDiffusionPipeline.from_pretrained(\n    \"stabilityai\/stable-video-diffusion-img2vid-xt\",\n    torch_dtype=torch.float16, variant=\"fp16\"\n)\npipe.enable_model_cpu_offload()\n\nimage = load_image(\"input.png\").resize((1024, 576))\nframes = pipe(image, decode_chunk_size=8, num_frames=25).frames[0]\nexport_to_video(frames, \"out.mp4\", fps=7)<\/code><\/pre>\n<p>For Wan 2.2 use <code>WanImageToVideoPipeline<\/code>; for LTX-Video use <code>LTXImageToVideoPipeline<\/code>. Check the model card for the exact class name \u2014 <code>diffusers<\/code> adds new pipelines each release.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"VRAM_and_hardware_honestly\"><\/span>VRAM and hardware, honestly<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>I2V is heavier than image generation because you&#8217;re denoising many frames of a temporal latent at once. Model cards and the diffusers repo report roughly:<\/p>\n<table>\n<thead>\n<tr>\n<th>Model<\/th>\n<th>Native VRAM (fp16)<\/th>\n<th>With CPU offload \/ quantization<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>SVD \/ SVD-XT (576\u00d71024)<\/td>\n<td>~16\u201320 GB<\/td>\n<td>~8\u201310 GB with <code>enable_model_cpu_offload()<\/code><\/td>\n<\/tr>\n<tr>\n<td>LTX-Video<\/td>\n<td>~12\u201324 GB depending on length<\/td>\n<td>Runs on 12 GB cards with offload<\/td>\n<\/tr>\n<tr>\n<td>Wan 2.1 I2V-1.3B<\/td>\n<td>~8\u201312 GB<\/td>\n<td>Runs on a 3060 12 GB<\/td>\n<\/tr>\n<tr>\n<td>Wan 2.2 I2V-A14B (720p)<\/td>\n<td>~24 GB+<\/td>\n<td>Fits on 16 GB with fp8\/GGUF community forks<\/td>\n<\/tr>\n<tr>\n<td>CogVideoX-5B-I2V<\/td>\n<td>~18\u201324 GB<\/td>\n<td>~5 GB with full offload (per THUDM card)<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>Numbers vary with resolution and frame count. For a general view of GPU memory scaling see the <a href=\"https:\/\/convly.ai\/llm-vram-calculator\/\">VRAM calculator<\/a> and <a href=\"https:\/\/convly.ai\/best-gpus-for-local-llms-2026\/\">best GPUs for local models<\/a> \u2014 the same cards (RTX 3090, 4090, 5090, A6000) are the sweet spot for video too.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Windows\"><\/span>Windows<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Install Python 3.10 or 3.11 from python.org, then the CUDA build of PyTorch from <a href=\"https:\/\/pytorch.org\/get-started\/locally\/\" rel=\"noopener\" target=\"_blank\">pytorch.org\/get-started\/locally<\/a> matching your NVIDIA driver. Create a venv, <code>pip install diffusers transformers accelerate<\/code>, and set <code>HF_HOME<\/code> to a drive with 50+ GB free \u2014 Wan and CogVideoX weights are large. Windows has no ROCm support; AMD users should use WSL2 or Linux. For a friendlier UI, <a href=\"https:\/\/github.com\/comfyanonymous\/ComfyUI\" rel=\"noopener\" target=\"_blank\">ComfyUI<\/a> has native workflows for SVD, Wan and LTX-Video.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"macOS\"><\/span>macOS<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Apple Silicon (M1\u2013M4) runs <code>diffusers<\/code> via MPS. Install PyTorch with MPS support, then pass <code>torch_dtype=torch.float16<\/code> and <code>.to(\"mps\")<\/code>. Realistically, SVD produces a 25-frame clip on an M2 Max in several minutes; Wan 2.2 A14B is impractical on anything below 64 GB unified memory. LTX-Video is currently the most usable I2V model on Mac because of its lower step count. Intel Macs have no working GPU path \u2014 use a hosted API instead.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Linux\"><\/span>Linux<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Linux is the primary target. Install the NVIDIA driver, CUDA 12.x runtime (bundled in PyTorch wheels), then <code>pip install diffusers transformers accelerate<\/code>. For multi-GPU or long clips use <code>accelerate config<\/code> and <code>pipe.enable_sequential_cpu_offload()<\/code>. <code>xformers<\/code> and <code>torch.compile()<\/code> give large speedups for SVD and Wan on Ampere\/Ada\/Hopper cards.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"When_to_skip_Hugging_Face_and_use_a_hosted_video_API\"><\/span>When to skip Hugging Face and use a hosted video API<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>If you need one short clip a day, a Space or a local RTX card is fine. If you need reliable throughput, per-second hosted APIs are usually cheaper than paying for an idle A100\/H100 endpoint:<\/p>\n<ul>\n<li><a href=\"https:\/\/convly.ai\/model\/wan-2-5\/\">Wan 2.5<\/a> \u2014 from $0.05\/sec (same lineage as the open Wan weights on the Hub)<\/li>\n<li><a href=\"https:\/\/convly.ai\/model\/sora-2\/\">Sora 2<\/a> \u2014 from $0.05\/sec; <a href=\"https:\/\/convly.ai\/model\/sora-2-pro\/\">Sora 2 Pro<\/a> from $0.15\/sec<\/li>\n<li><a href=\"https:\/\/convly.ai\/model\/veo-3-1\/\">Veo 3.1<\/a> \u2014 from $0.05\/sec<\/li>\n<li><a href=\"https:\/\/convly.ai\/model\/kling-2-5-turbo-pro\/\">Kling 2.5 Turbo Pro<\/a> \u2014 from $0.07\/sec<\/li>\n<\/ul>\n<p>At $0.05 per second, 1,000 five-second clips cost $250; renting an H100 at typical cloud rates for the same throughput usually costs more. Use the <a href=\"https:\/\/convly.ai\/ai-api-cost-calculator\/\">API cost calculator<\/a> and browse the <a href=\"https:\/\/convly.ai\/models\/\">models database<\/a> for current pricing.<\/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 there a single &#8220;Hugging Face image to video&#8221; model I should default to?<\/h3>\n<p>No single default. For open weights and permissive licensing, Wan 2.2 I2V-A14B currently produces the best 720p output but wants 24 GB+ of VRAM. On 12 GB cards, LTX-Video or Wan 2.1 I2V-1.3B are the practical choices. SVD-XT is older but very well documented.<\/p>\n<h3>Can I use these models commercially?<\/h3>\n<p>Depends on the checkpoint. Wan 2.1\/2.2, CogVideoX code and LTX-Video are permissive but each has its own terms; Stable Video Diffusion&#8217;s weights ship under a Stability community license with a separate commercial tier. Always read the model card&#8217;s License section before shipping anything.<\/p>\n<h3>Why does my output flicker or look like a slideshow?<\/h3>\n<p>Usually one of: too few inference steps (raise <code>num_inference_steps<\/code>), <code>motion_bucket_id<\/code>\/<code>guidance_scale<\/code> set too low for SVD, or an input image whose aspect ratio doesn&#8217;t match the model&#8217;s training resolution. Match the model&#8217;s native resolution (e.g. 1024\u00d7576 for SVD-XT) before generating.<\/p>\n<h3>Can I run image-to-video on a laptop with 8 GB VRAM?<\/h3>\n<p>Only with the smallest models and aggressive offloading. Wan 2.1 I2V-1.3B and quantized LTX-Video builds can run, but expect minutes per clip. For anything larger a hosted API or a Space will be faster and cheaper than fighting OOM errors.<\/p>\n<h3>Does Ollama support image-to-video?<\/h3>\n<p>No. Ollama is focused on text and multimodal LLMs, not diffusion video pipelines. See the <a href=\"https:\/\/convly.ai\/what-is-ollama-complete-guide-2026\/\">Ollama guide<\/a> for what it does cover; for I2V, stick with <code>diffusers<\/code> or ComfyUI.<\/p>\n<h3>How long does a single clip take on a consumer GPU?<\/h3>\n<p>Order of magnitude: SVD-XT on an RTX 4090 produces a 25-frame clip in well under a minute at default settings; the same model with <code>enable_model_cpu_offload()<\/code> on a 12 GB card is several times slower. Wan 2.2 A14B at 720p is substantially heavier. Exact numbers depend on step count, resolution and PyTorch version, so treat any single benchmark you read as approximate.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Hugging Face hosts image-to-video (I2V) models you can run three ways: hosted Spaces demos, the Inference API\/Endpoints, or locally with [\u2026]<\/p>\n","protected":false},"author":1,"featured_media":2726,"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":[],"class_list":["post-2725","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-tutorials"],"_links":{"self":[{"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/posts\/2725","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/comments?post=2725"}],"version-history":[{"count":1,"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/posts\/2725\/revisions"}],"predecessor-version":[{"id":2727,"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/posts\/2725\/revisions\/2727"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/media\/2726"}],"wp:attachment":[{"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/media?parent=2725"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/categories?post=2725"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/convly.ai\/fr\/wp-json\/wp\/v2\/tags?post=2725"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}