{"id":16846,"date":"2025-10-20T09:04:45","date_gmt":"2025-10-20T06:04:45","guid":{"rendered":"https:\/\/rediteq.fi\/?p=16846"},"modified":"2025-10-20T09:06:57","modified_gmt":"2025-10-20T06:06:57","slug":"how-to-solve-the-challenges-of-narrative-ai-video","status":"publish","type":"post","link":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video","title":{"rendered":"How to Solve the Challenges of Narrative AI Video"},"content":{"rendered":"\r\n\r\n\r\n\r\n\r\n\r\nYou\u2019re probably reading this blog because you got excited about creating striking short AI video clips and wanted to move on to longer, narrative videos. However, you noticed that at this point the challenges increased, and the results were no longer as satisfying or as intended. You\u2019ve seen professionally made long-form AI videos, so you know this isn\u2019t an impossible task.\r\n\r\nIn narrative stories, videos consist of scenes that are connected not only by the plot but also by a visual throughline\u2014such as recurring characters, devices, style, etc. It could be an entertaining animation or, for example, a series of first-aid instructions where the same instructor demonstrates procedures for different situations. You want the animated character to look exactly the same throughout the video and the first-aid instructor to be the same person, dressed the same way in different scenes.\r\n\r\nWhen you made your first AI video experiments, you likely started with text-to-video features, which let you wow yourself and others with an imaginative character or surreal action. You may even have used a well-known public figure in the video. The situation is very different when you want to create an adventure for your own character or instructional content from a first-aid instructor. When you make a single video clip, it\u2019s enough that the character\u2019s appearance suits that scene. If you try to create a follow-up with the same description, the AI will often generate the character looking different. If, in your experiments, you used a public figure, the AI recognizes the person and you can produce another scene with the same character; but if you\u2019re creating videos for professional use, you generally can\u2019t use such material for legal (e.g., copyright) reasons.\r\n

1: Always Use a Starting Image<\/b><\/span><\/h2>\r\nText-to-video is a rapidly evolving and exciting way to create videos, but when your goal is to produce videos that match your own imagination and objectives, you should practically always use a starting image<\/b>\u2014or, in some cases, a reference image<\/b>, which some AI models support. Using a starting image is called image-to-video<\/b>. In this workflow, the first frame of the video is your starting image, and you direct events and content with text. You\u2019ll need one starting image for each scene. You can create starting images in AI services, or they can be any pictures\u2014like ones taken on a phone.\r\n

2: Choose the Image Model to Suit Your Needs<\/b><\/span><\/h2>\r\nCreating images and creating videos differ quite a lot technically. When aiming for high quality, the core challenges in generating them are quite different. The same provider may offer both at a sufficiently high level, but the best models for image generation are often not the best or most suitable for video generation\u2014and vice versa. There are many products and AI models dedicated solely to image creation. Their emphases and aesthetic baselines vary significantly, which directly affects how your video will ultimately look. Let\u2019s briefly compare two widely used image services: Midjourney and OpenAI\u2019s DALL\u00b7E.\r\n\r\nMidjourney<\/b> is known for an almost artistic approach. It produces images whose lighting, colors, and composition can resemble a carefully crafted advertising photo. If your goal is a visually coherent animation style where the world and characters feel like they come from the same universe, Midjourney is an excellent choice. It\u2019s especially good for concept images, backgrounds, and character design.\r\n\r\nDALL\u00b7E<\/b> (especially the newer versions like DALL\u00b7E 3) approaches images semantically\u2014that is, through logical content. Where Midjourney leans on mood and aesthetics, DALL\u00b7E aims to understand what should appear in the image and why. As a result, DALL\u00b7E is often better when you need clearly defined content, such as product use cases, instructions, or training materials.\r\n

3: Choose the Video Model to Suit Your Needs<\/b><\/span><\/h2>\r\nThere are many options for creating videos, and progress is very fast these days. When selecting a model, consider whether you primarily want visually striking individual scenes or a longer story where characters and environments remain consistent. Also consider whether you want the video to include built-in audio. Price naturally matters as well. Costs vary a lot, and the price per usable second of video ranges from a few cents to tens of cents. In longer productions, price matters, and you must account for the fact that some generated material will be cut out. Here is a brief comparison of some well-known AI video models:\r\n\r\nRunway Gen-3<\/b> is an excellent choice when you want creative, dynamic, cinematic clips. It\u2019s designed for rapid ideation and visual experimentation, not necessarily for maintaining long-form continuity. A challenge is consistency in characters and environments. A benefit, however, is support for reference images<\/b>, which helps you keep the same characters across clips.\r\n\r\nOpenAI\u2019s Sora 2<\/b> is designed for generating long, continuous, and natural video. It doesn\u2019t only understand text prompts but also movement, physics, and cause-and-effect. For creating realistic, story-driven video, Sora 2 is currently among the best options. Sora 2 also generates audio for videos.\r\n\r\nCompeting with Sora (and Runway) is Google\u2019s VEO 3<\/b>, which also offers audio in videos. The latest version, VEO 3.1<\/b>, emphasizes credible, physics-abiding motion realism. It is designed for multi-scene storytelling and supports, among other things, the use of reference images.\r\n\r\nThese products are by no means the only viable options. For example, Kling AI<\/b> from the Chinese tech company Kuaishou challenges other top-tier video models with a rapid development cycle. At the time of writing, the newest version is Kling 2.5 Turbo<\/b>, which is a credible alternative. Synthesia<\/b> offers a different approach: it\u2019s not a creative \u201cvideo generator\u201d but a content production platform that uses ready-made virtual characters (talking avatars). It can be a good choice when your goal is a clearly structured, informative video\u2014such as an instruction, onboarding, or presentation. Its drawback is visual limitation: the style and motion aren\u2019t especially creative but rather standardized for business use.\r\n

4: Design the Framework First<\/b><\/span><\/h2>\r\nMany are disappointed when the second scene looks completely different from the first. This is because AI models don\u2019t yet \u201cremember\u201d previous prompts\u2014they generate each image and video separately. This current characteristic of AI video generation must be accounted for in planning.\r\n\r\nConsistency arises from design, not chance. Use an approach analogous to the film industry\u2019s pre-production<\/b>: first define the visual language, and only then \u201cshoot.\u201d In AI video production, this means controlled use of reference images and starting images.\r\n\r\nDuring planning, you write the script and assign roles to characters. You must consider what is presented and how. To answer the \u201cwhat,\u201d you design scenes based on content. To convey the content or story, you must plan who or what appears in each scene. In practice, your framework should include a sequence of scenes and the starting images for those scenes.\r\n

5: Build Videos from Small Pieces<\/b><\/span><\/h2>\r\nThe biggest challenge in video production today is manageability. As with many other tasks and projects, manageability improves when you break a large whole into small, controllable parts. In video production, this means creating the video in pieces short enough that not a single error \u201cfits\u201d into one piece. This often means that within the piece you\u2019re working on, only one movement or action happens at a time. When the story has progressed as intended, you move on to the next piece, and then the next. Today\u2019s AI services typically let you create short clips of about 3\u201315 seconds at a time. In challenging parts, progress can be much slower\u2014you may only advance by a second or two at a time to achieve sufficient quality. As models improve, you\u2019ll be able to generate longer ready-to-use clips in one go. Even in the future, though, larger wholes will be built from smaller pieces, so it\u2019s worth getting comfortable with this approach.\r\n

6: How to Gain More Control<\/b><\/span><\/h2>\r\nAs described in the previous section, you gain control by breaking your project into small parts. That raises the question of how these parts work together. If you want characters, lighting, and style to remain consistent throughout the video, you need tools and methods that guide the AI consistently. This is where process-guiding tools and, for example, reference images become especially important.\r\n\r\nA key method is to structure your planning process<\/b> before generating any video. In practice, this means you don\u2019t start with AI but with paper or a digital tool that supports scene planning. That way you won\u2019t get stuck in a dead end and can craft the story using the techniques available.\r\n\r\nReference images<\/b> are a useful feature that helps keep scenes consistent. Unfortunately, many AI models still do not support reference images today. You can think of reference images as a control system for the AI\u2014they tell the model what the world should look like before it starts generating motion. As a substitute for reference images, you can use starting images.\r\n\r\nUsing reference images and planning tools restores what has always been crucial in film production: visual consistency and narrative rhythm. AI video production currently resembles directing<\/b> more than mere \u201cgeneration.\u201d Every prompt, reference image, and parameter tweak is a directorial decision that affects the outcome.\r\n

7: Typical AI Errors and Shortcomings<\/b><\/span><\/h2>\r\nAI videos are under intense development, and quality is improving rapidly. There is still plenty of room for improvement, however. Almost everyone experimenting with these tools has encountered the following technical issues:\r\n\r\nObject distortions:<\/b>\r\n<\/b> Hands, fingers, clothing edges, and thin objects are easily distorted because the models don\u2019t yet fully handle natural 3D structure and contact points.\r\nFix:<\/b>\r\n<\/b> \u2013 Shorten clip length. The longer the generation, the higher the risk of distortion in motion.\r\n\u2013 Use the image-to-video<\/b> method\u2014it anchors form more stably than text-to-video alone.\r\n\u2013 Avoid complex actions (e.g., \u201cthe character picks up a mug from the table and waves\u201d)\u2014split them into two shots.\r\n\r\nLack of memory:<\/b>\r\n<\/b> The AI doesn\u2019t \u201cremember\u201d previous takes; it recreates each image and video. That\u2019s why the same character may look slightly different in each clip\u2014different hair, facial features, or even clothing.\r\nFix:<\/b>\r\n<\/b> \u2013 Use a reference image or reference video in every generation step.\r\n\u2013 Create a character bank<\/b>.\r\n\u2013 Add detailed prompt constraints: \u201csame woman, same outfit, same hairstyle as previous shot.\u201d\r\n

8: The Costs of AI Video Production<\/b><\/span><\/h2>\r\nOne of the biggest promises of AI video has been cost-effectiveness. And it\u2019s true\u2014you can now create a video with AI that would easily have cost thousands of euros (or much more) with traditional methods, depending on the content. At the same time, it\u2019s good to understand that AI videos aren\u2019t \u201cfree,\u201d nor always very cheap. Costs are distributed differently: some relate to compute, some to service licenses, and some to your own time and planning.\r\n\r\nAs in almost all projects, planning<\/b> is the key to cost savings here as well. Always start from the \u201cbig picture\u201d: think through and describe your goals. Write the script and plan the scenes. Only once the plan is ready should you proceed to the actual production phase\u2014in AI video, the generation itself. You may be able to create a first draft<\/b> with a cheaper model and at a lower resolution to validate your ideas.\r\n\r\nPlanning takes time and thus costs money\u2014but it\u2019s also clear that if you start generating videos without a plan, you\u2019ll create a lot of throwaway footage, and every second costs<\/b>. In practice, you can create the plan either before generation or iteratively during it. The more experienced you are, the better you can invest in planning and thereby achieve higher quality at lower cost.\r\n\r\n","protected":false},"excerpt":{"rendered":"

You\u2019re probably reading this blog because you got excited about creating striking short AI video clips and wanted to move…<\/p>\n","protected":false},"author":8,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"sfsi_plus_gutenberg_text_before_share":"","sfsi_plus_gutenberg_show_text_before_share":"","sfsi_plus_gutenberg_icon_type":"","sfsi_plus_gutenberg_icon_alignemt":"","sfsi_plus_gutenburg_max_per_row":"","footnotes":""},"categories":[15],"tags":[],"class_list":{"0":"post-16846","1":"post","2":"type-post","3":"status-publish","4":"format-standard","6":"category-ajankohtaiset-en"},"featured_image_src":null,"author_info":{"display_name":"S","author_link":"https:\/\/rediteq.fi\/en\/author\/raakel"},"yoast_head":"\nHow to Solve the Challenges of Narrative AI Video - Rediteq Oy<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"How to Solve the Challenges of Narrative AI Video - Rediteq Oy\" \/>\n<meta property=\"og:description\" content=\"You\u2019re probably reading this blog because you got excited about creating striking short AI video clips and wanted to move...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video\" \/>\n<meta property=\"og:site_name\" content=\"Rediteq Oy\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/Rediteq\/?fref=ts\" \/>\n<meta property=\"article:published_time\" content=\"2025-10-20T06:04:45+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2025-10-20T06:06:57+00:00\" \/>\n<meta name=\"author\" content=\"S\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"S\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"10 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video#article\",\"isPartOf\":{\"@id\":\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video\"},\"author\":{\"name\":\"S\",\"@id\":\"https:\/\/rediteq.fi\/#\/schema\/person\/3c67220d168b384a29ed627d9a2ec6ca\"},\"headline\":\"How to Solve the Challenges of Narrative AI Video\",\"datePublished\":\"2025-10-20T06:04:45+00:00\",\"dateModified\":\"2025-10-20T06:06:57+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video\"},\"wordCount\":1947,\"publisher\":{\"@id\":\"https:\/\/rediteq.fi\/#organization\"},\"articleSection\":[\"Ajankohtaiset\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video\",\"url\":\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video\",\"name\":\"How to Solve the Challenges of Narrative AI Video - Rediteq Oy\",\"isPartOf\":{\"@id\":\"https:\/\/rediteq.fi\/#website\"},\"datePublished\":\"2025-10-20T06:04:45+00:00\",\"dateModified\":\"2025-10-20T06:06:57+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/rediteq.fi\/en\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"How to Solve the Challenges of Narrative AI Video\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/rediteq.fi\/#website\",\"url\":\"https:\/\/rediteq.fi\/\",\"name\":\"Rediteq Oy\",\"description\":\"Koulutusratkaisut yrityksille ja oppilaitoksille\",\"publisher\":{\"@id\":\"https:\/\/rediteq.fi\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/rediteq.fi\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/rediteq.fi\/#organization\",\"name\":\"Rediteq Oy\",\"url\":\"https:\/\/rediteq.fi\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/rediteq.fi\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/rediteq.fi\/wp-content\/uploads\/2017\/02\/rediteq_logo_png_600px.png\",\"contentUrl\":\"https:\/\/rediteq.fi\/wp-content\/uploads\/2017\/02\/rediteq_logo_png_600px.png\",\"width\":600,\"height\":140,\"caption\":\"Rediteq Oy\"},\"image\":{\"@id\":\"https:\/\/rediteq.fi\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.facebook.com\/Rediteq\/?fref=ts\",\"https:\/\/www.instagram.com\/rediteqfi\/\",\"https:\/\/www.youtube.com\/channel\/UChdWBDfytwfkNSFYdMv33MA\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/rediteq.fi\/#\/schema\/person\/3c67220d168b384a29ed627d9a2ec6ca\",\"name\":\"S\",\"url\":\"https:\/\/rediteq.fi\/en\/author\/raakel\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"How to Solve the Challenges of Narrative AI Video - Rediteq Oy","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video","og_locale":"en_US","og_type":"article","og_title":"How to Solve the Challenges of Narrative AI Video - Rediteq Oy","og_description":"You\u2019re probably reading this blog because you got excited about creating striking short AI video clips and wanted to move...","og_url":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video","og_site_name":"Rediteq Oy","article_publisher":"https:\/\/www.facebook.com\/Rediteq\/?fref=ts","article_published_time":"2025-10-20T06:04:45+00:00","article_modified_time":"2025-10-20T06:06:57+00:00","author":"S","twitter_card":"summary_large_image","twitter_misc":{"Written by":"S","Est. reading time":"10 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video#article","isPartOf":{"@id":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video"},"author":{"name":"S","@id":"https:\/\/rediteq.fi\/#\/schema\/person\/3c67220d168b384a29ed627d9a2ec6ca"},"headline":"How to Solve the Challenges of Narrative AI Video","datePublished":"2025-10-20T06:04:45+00:00","dateModified":"2025-10-20T06:06:57+00:00","mainEntityOfPage":{"@id":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video"},"wordCount":1947,"publisher":{"@id":"https:\/\/rediteq.fi\/#organization"},"articleSection":["Ajankohtaiset"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video","url":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video","name":"How to Solve the Challenges of Narrative AI Video - Rediteq Oy","isPartOf":{"@id":"https:\/\/rediteq.fi\/#website"},"datePublished":"2025-10-20T06:04:45+00:00","dateModified":"2025-10-20T06:06:57+00:00","breadcrumb":{"@id":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/rediteq.fi\/en\/how-to-solve-the-challenges-of-narrative-ai-video#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/rediteq.fi\/en"},{"@type":"ListItem","position":2,"name":"How to Solve the Challenges of Narrative AI Video"}]},{"@type":"WebSite","@id":"https:\/\/rediteq.fi\/#website","url":"https:\/\/rediteq.fi\/","name":"Rediteq Oy","description":"Koulutusratkaisut yrityksille ja oppilaitoksille","publisher":{"@id":"https:\/\/rediteq.fi\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/rediteq.fi\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/rediteq.fi\/#organization","name":"Rediteq Oy","url":"https:\/\/rediteq.fi\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/rediteq.fi\/#\/schema\/logo\/image\/","url":"https:\/\/rediteq.fi\/wp-content\/uploads\/2017\/02\/rediteq_logo_png_600px.png","contentUrl":"https:\/\/rediteq.fi\/wp-content\/uploads\/2017\/02\/rediteq_logo_png_600px.png","width":600,"height":140,"caption":"Rediteq Oy"},"image":{"@id":"https:\/\/rediteq.fi\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/Rediteq\/?fref=ts","https:\/\/www.instagram.com\/rediteqfi\/","https:\/\/www.youtube.com\/channel\/UChdWBDfytwfkNSFYdMv33MA"]},{"@type":"Person","@id":"https:\/\/rediteq.fi\/#\/schema\/person\/3c67220d168b384a29ed627d9a2ec6ca","name":"S","url":"https:\/\/rediteq.fi\/en\/author\/raakel"}]}},"_links":{"self":[{"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/posts\/16846","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/comments?post=16846"}],"version-history":[{"count":2,"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/posts\/16846\/revisions"}],"predecessor-version":[{"id":16848,"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/posts\/16846\/revisions\/16848"}],"wp:attachment":[{"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/media?parent=16846"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/categories?post=16846"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/rediteq.fi\/en\/wp-json\/wp\/v2\/tags?post=16846"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}