<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Kreavo</title>
    <description>The latest articles on DEV Community by Kreavo (@kreavo).</description>
    <link>https://dev.to/kreavo</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4215989%2F3fc0a083-2a27-45c8-bffc-893b9b86cb74.png</url>
      <title>DEV Community: Kreavo</title>
      <link>https://dev.to/kreavo</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/kreavo"/>
    <language>en</language>
    <item>
      <title>Generate AI UGC videos from Claude, ChatGPT or Cursor with MCP</title>
      <dc:creator>Kreavo</dc:creator>
      <pubDate>Thu, 05 Nov 2026 11:38:32 +0000</pubDate>
      <link>https://dev.to/kreavo/generate-ai-ugc-videos-from-claude-chatgpt-or-cursor-with-mcp-23l5</link>
      <guid>https://dev.to/kreavo/generate-ai-ugc-videos-from-claude-chatgpt-or-cursor-with-mcp-23l5</guid>
      <description>&lt;p&gt;&lt;strong&gt;Short answer:&lt;/strong&gt; add &lt;code&gt;https://mcp.kreavo.ai/mcp&lt;/code&gt; as a remote MCP server in your client, sign in once through OAuth, and ask for a video. The model calls &lt;code&gt;kreavo_generate_video&lt;/code&gt;, gets back an ID, then polls &lt;code&gt;kreavo_check_video_status&lt;/code&gt; until a video URL comes back. On the default settings (5 s, 720p) that costs 140 credits, or $1.40.&lt;/p&gt;

&lt;h2&gt;
  
  
  MCP in three lines
&lt;/h2&gt;

&lt;p&gt;The Model Context Protocol is a standard way for an LLM client to discover and call tools exposed by a server. The client lists the tools, the model decides when to call them, and the server does the work. Here the server is remote: it speaks Streamable HTTP and uses OAuth, so you don't install a package or paste an API key.&lt;/p&gt;

&lt;h2&gt;
  
  
  Setup
&lt;/h2&gt;

&lt;h3&gt;
  
  
  Claude Code
&lt;/h3&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;claude mcp add &lt;span class="nt"&gt;--transport&lt;/span&gt; http kreavo https://mcp.kreavo.ai/mcp
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Or commit it to a project-level &lt;code&gt;.mcp.json&lt;/code&gt; (Claude Code asks you to approve project servers the first time):&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"mcpServers"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"kreavo"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"type"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"http"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"url"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"https://mcp.kreavo.ai/mcp"&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h3&gt;
  
  
  Claude Desktop / claude.ai
&lt;/h3&gt;

&lt;ol&gt;
&lt;li&gt;Settings → Connectors&lt;/li&gt;
&lt;li&gt;"Add custom connector"&lt;/li&gt;
&lt;li&gt;Name &lt;code&gt;Kreavo&lt;/code&gt;, URL &lt;code&gt;https://mcp.kreavo.ai/mcp&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;Connect, then sign in with your Kreavo account&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;On Team and Enterprise plans, an owner adds the connector first from Admin settings → Connectors.&lt;/p&gt;

&lt;h3&gt;
  
  
  ChatGPT
&lt;/h3&gt;

&lt;p&gt;There is no listed Kreavo plugin in ChatGPT yet, so you add the server by hand. At the time of writing, that means going to the Plugins page (chatgpt.com/plugins), clicking the plus button, choosing "Add custom MCP server", entering &lt;code&gt;https://mcp.kreavo.ai/mcp&lt;/code&gt; as the public endpoint and picking OAuth for authentication. Earlier versions put this under Settings → Connectors, with developer mode turned on, and workspace policies can require an admin to allow custom servers. ChatGPT renames these menus fairly often, so look for "custom MCP server" or "Connectors" wherever the current screen puts them.&lt;/p&gt;

&lt;p&gt;There are a few differences when you use it from ChatGPT. Tools that spend credits are flagged as destructive, so ChatGPT asks you to confirm before each paid generation. The &lt;code&gt;model&lt;/code&gt; parameter is limited to a shorter list. Requests for sexual content or nudity are blocked before anything is created or charged. Images you attach in the chat are accepted as inputs.&lt;/p&gt;

&lt;h3&gt;
  
  
  Cursor
&lt;/h3&gt;

&lt;p&gt;&lt;code&gt;~/.cursor/mcp.json&lt;/code&gt; (global) or &lt;code&gt;.cursor/mcp.json&lt;/code&gt; (project):&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"mcpServers"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"kreavo"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"url"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"https://mcp.kreavo.ai/mcp"&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h3&gt;
  
  
  VS Code
&lt;/h3&gt;

&lt;p&gt;&lt;code&gt;.vscode/mcp.json&lt;/code&gt; in the workspace, or your user &lt;code&gt;mcp.json&lt;/code&gt; (command "MCP: Open User Configuration"):&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight json"&gt;&lt;code&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"servers"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"kreavo"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"type"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"http"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="nl"&gt;"url"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="s2"&gt;"https://mcp.kreavo.ai/mcp"&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt; &lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="w"&gt;
&lt;/span&gt;&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h3&gt;
  
  
  What the login looks like
&lt;/h3&gt;

&lt;p&gt;The first time the client calls the server, it gets a &lt;code&gt;401&lt;/code&gt; that points it to the OAuth metadata. It then opens your browser on a consent screen titled "Connect Claude" (or ChatGPT, Cursor, VS Code…), with the account email and "Authorize" / "Cancel" buttons. If you aren't signed in to Kreavo in that browser, you sign in first and land back on that screen. The flow uses PKCE (S256) with dynamic client registration. Access tokens last 7 days. Refresh tokens last 90 days and change each time they're used. If the confirmation page opens in a different browser (for example from a mail app), it tells you to finish in the browser that started the flow. That's on purpose: the request is tied to that browser.&lt;/p&gt;

&lt;h2&gt;
  
  
  A real conversation, step by step
&lt;/h2&gt;

&lt;p&gt;Prompt in Claude Desktop, with a link to the product photo (here, the image URL from the product page):&lt;/p&gt;

&lt;blockquote&gt;
&lt;p&gt;Make a 5-second vertical UGC-style clip of the serum bottle in this photo: https://… — a woman picks it up from a bathroom shelf, smiles at the camera, natural daylight. Check what it costs first.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;The tool calls, shortened (IDs and URLs removed):&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;→ kreavo_list_models
← table of models with a "Credits" column (seedance-2-fast: 140 for 5 s at 720p)

→ kreavo_generate_video {
    "prompt": "A woman picks up the serum bottle from a bathroom shelf, smiles at the camera, handheld phone footage, natural daylight",
    "image_url": "&amp;lt;product photo URL&amp;gt;",
    "model": "seedance-2-fast",
    "duration": 5,
    "resolution": "720p",
    "aspect_ratio": "9:16"
  }
← Video ID: &amp;lt;id&amp;gt;. Use kreavo_check_video_status with task_id=&amp;lt;id&amp;gt;

→ kreavo_check_video_status { "task_id": "&amp;lt;id&amp;gt;" }
← Video Still Processing ⏳ … Try again in 15-30 seconds

→ kreavo_check_video_status { "task_id": "&amp;lt;id&amp;gt;" }
← Video Ready ✅ &amp;lt;video URL&amp;gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The video also shows up in the account's gallery on kreavo.ai, so you don't have to save the URL yourself.&lt;/p&gt;

&lt;p&gt;Why a URL and not an attached photo? In Claude and Cursor, the model can see an image you attach, but it can't hand that file to a tool. &lt;code&gt;kreavo_upload_image&lt;/code&gt; exists for local files: it takes the image as base64 text and returns a public URL. Since the model has to write that whole base64 string into the call, it's only practical for small files, for example from Claude Code. For anything bigger, give a public https link to the image. ChatGPT is the exception: images attached in the chat are passed to the tools as files.&lt;/p&gt;

&lt;h2&gt;
  
  
  How the async part works
&lt;/h2&gt;

&lt;p&gt;Video generation takes minutes, which is longer than you'd want one tool call to stay open. So the three video tools (&lt;code&gt;kreavo_generate_video&lt;/code&gt;, &lt;code&gt;kreavo_generate_product_video&lt;/code&gt;, &lt;code&gt;kreavo_control_motion&lt;/code&gt;) return an ID right away along with an explicit instruction to poll. The model reads that instruction and calls &lt;code&gt;kreavo_check_video_status&lt;/code&gt;, which answers in one of three ways:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Still Processing&lt;/strong&gt;: wait 15–30 s and call again.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Ready&lt;/strong&gt;: the video URL, sent as text. MCP has no video content type, so you get a link and not an inline player.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Failed&lt;/strong&gt;: a short reason tag such as &lt;code&gt;moderation:content_policy&lt;/code&gt; (every video failure comes back in this &lt;code&gt;moderation:&amp;lt;reason&amp;gt;&lt;/code&gt; form). Failed generations are refunded.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Most image models answer synchronously. The image comes back inline if it's 950 KB or less, otherwise you get just the URL. &lt;code&gt;gpt-image-2&lt;/code&gt; is the exception: it returns a task ID, and you poll &lt;code&gt;kreavo_check_image_status&lt;/code&gt; the same way.&lt;/p&gt;

&lt;p&gt;If your agent loops by itself, give it a rule like "wait about 20 seconds between status checks". Otherwise some clients poll too eagerly and burn through the rate limit.&lt;/p&gt;

&lt;h2&gt;
  
  
  Costs and limits, honestly
&lt;/h2&gt;

&lt;p&gt;1 credit = $0.01. These are the live prices from the billing functions as of November 2026:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Video model&lt;/th&gt;
&lt;th&gt;5 s, 720p&lt;/th&gt;
&lt;th&gt;Notes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;
&lt;code&gt;seedance-2-fast&lt;/code&gt; (default)&lt;/td&gt;
&lt;td&gt;140&lt;/td&gt;
&lt;td&gt;280 for 10 s&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;seedance-2-5&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;175&lt;/td&gt;
&lt;td&gt;350 for 10 s, 245 at 1080p&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;kling-3-0-std&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;140&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;kling-3-0-pro&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;187&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;veo3.1-fast&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;200&lt;/td&gt;
&lt;td&gt;fixed 8 s&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;veo3.1-quality&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;500&lt;/td&gt;
&lt;td&gt;fixed 8 s&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;kreavo-motion-fast&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;55&lt;/td&gt;
&lt;td&gt;5 or 8 s, 70 for 8 s&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Image model&lt;/th&gt;
&lt;th&gt;Credits&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;
&lt;code&gt;nano-banana-2&lt;/code&gt; (default)&lt;/td&gt;
&lt;td&gt;70 at 4K (its default), 55 at 2K&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;nano-banana-2-1&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;25 at 2K, 40 at 4K&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;nano-banana-pro&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;60 (2K)&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;gpt-image-2&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;8 (async)&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;&lt;code&gt;flux-schnell&lt;/code&gt;&lt;/td&gt;
&lt;td&gt;4&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Other tools: &lt;code&gt;kreavo_create_character&lt;/code&gt; costs 50 credits, &lt;code&gt;kreavo_virtual_tryon&lt;/code&gt; costs 55 (paid plans only), and &lt;code&gt;kreavo_control_motion&lt;/code&gt; costs 9 credits per second at 720p or 14 per second at 1080p, based on the length of the reference video (3 to 30 s, rounded up to the next second).&lt;/p&gt;

&lt;p&gt;Things to know before you give an agent this server:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;No price prompt before a charge.&lt;/strong&gt; The generation tools spend credits straight away. Every paid tool's description says it spends credits and points to &lt;code&gt;kreavo_list_models&lt;/code&gt;, but checking the price is up to the model, or up to you in the prompt. ChatGPT is the exception because it asks for confirmation.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Video length:&lt;/strong&gt; 4–15 s for most models. Seedance 2.5 goes up to 30 s, Kling 3.0 starts at 5 s, Kreavo Motion is 5 or 8 s, and Veo 3.1 is fixed at 8 s.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Rate limit:&lt;/strong&gt; 120 requests per minute per account. Above that you get &lt;code&gt;429&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Uploads:&lt;/strong&gt; PNG, JPEG, WebP or GIF, up to 10 MB each, 30 uploads per hour, 200 files and 200 MB per day.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Sessions:&lt;/strong&gt; closed after 24 h without activity. Clients reconnect on their own.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Not enough credits:&lt;/strong&gt; the tool returns an error that says so and nothing is charged. Video needs a paid plan or the $1 trial.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Three workflows that work well
&lt;/h2&gt;

&lt;p&gt;&lt;strong&gt;1. Ten ad angles for one Shopify product.&lt;/strong&gt; Give the agent the product page and ask for 10 hooks: problem/solution, unboxing, before/after, "POV: you finally found…", and so on. It takes the product photo URL from the page, then calls &lt;code&gt;kreavo_generate_video&lt;/code&gt; ten times with different prompts and polls each ID. At 5 s / 720p on &lt;code&gt;seedance-2-fast&lt;/code&gt; that's 1,400 credits ($14) for ten clips you can test as ads.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;2. A consistent AI influencer.&lt;/strong&gt; &lt;code&gt;kreavo_create_character&lt;/code&gt; (50 credits) once, &lt;code&gt;kreavo_list_scenes&lt;/code&gt; to choose sets (studio, lifestyle, travel…), &lt;code&gt;kreavo_generate_photoshoot&lt;/code&gt; for stills in each scene, then &lt;code&gt;kreavo_generate_product_video&lt;/code&gt; with the &lt;code&gt;character_id&lt;/code&gt;, a product image and a script. The same face shows up in every asset, which you can't count on when each clip is prompted from scratch.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;3. Assets from your editor.&lt;/strong&gt; In Cursor or VS Code, have the agent read a product JSON in your repo and generate a set of images with &lt;code&gt;gpt-image-2&lt;/code&gt; (8 credits each), or catalog shots with &lt;code&gt;kreavo_virtual_tryon&lt;/code&gt;, while you work on the storefront code. The URLs come back into the conversation, so the agent can write them straight into your fixtures.&lt;/p&gt;

&lt;h2&gt;
  
  
  Wrapping up
&lt;/h2&gt;

&lt;p&gt;MCP turns "make me a video" into a few typed tool calls: generate, then poll. Your client handles auth and decides when to call each tool. The tools are the same in Claude, ChatGPT, Cursor and VS Code (ChatGPT adds the checks described above), so whatever prompts and agent loops you write carry over from one client to another.&lt;/p&gt;

&lt;p&gt;If you want to try it, the setup page is &lt;a href="https://kreavo.ai/mcp?utm_source=devto&amp;amp;utm_medium=article&amp;amp;utm_campaign=mcp" rel="noopener noreferrer"&gt;kreavo.ai/mcp&lt;/a&gt;. New accounts get 150 free credits for images, and video starts with a $1 trial.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>mcp</category>
      <category>tutorial</category>
      <category>productivity</category>
    </item>
    <item>
      <title>Creatify Alternative: How to Choose an AI UGC Ad Tool for a Physical Product (2026)</title>
      <dc:creator>Kreavo</dc:creator>
      <pubDate>Mon, 02 Nov 2026 12:05:21 +0000</pubDate>
      <link>https://dev.to/kreavo/creatify-alternative-how-to-choose-an-ai-ugc-ad-tool-for-a-physical-product-2026-lm6</link>
      <guid>https://dev.to/kreavo/creatify-alternative-how-to-choose-an-ai-ugc-ad-tool-for-a-physical-product-2026-lm6</guid>
      <description>&lt;blockquote&gt;
&lt;p&gt;&lt;strong&gt;Short answer:&lt;/strong&gt; if you are looking for a Creatify alternative, first decide what kind of ad you need more of. If it is many variations from a product URL, Creatify is built for exactly that and switching may not help. If it is one recognizable AI creator showing your real product, compare tools on five things: whether the product is really in the shot, whether the face stays the same, cost per usable clip, which video models you can choose, and what happens when a render fails.&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;Disclosure: we build Kreavo, one of the alternatives. We will describe Creatify the way it describes itself, and say where switching is not worth it.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why people look for a Creatify alternative
&lt;/h2&gt;

&lt;p&gt;Creatify is known for turning a product URL into ad videos with AI avatars and generating many variations in batch. Its own site lists features such as URL-to-video, ad clone, product video and an AI influencer generator, plus use-case pages for ecommerce, DTC and app ads. For finding a message that works, volume is a legitimate strategy.&lt;/p&gt;

&lt;p&gt;The reasons teams look around are usually about fit, not quality:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;The product.&lt;/strong&gt; A physical product brand needs the item in hand, readable label included. A URL-to-ad flow works from what is on the page; some brands want more control over the scene.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The face.&lt;/strong&gt; Rotating avatars are good for testing messages. Some brands want one AI creator that viewers recognize from ad to ad.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The budget shape.&lt;/strong&gt; A subscription sized for volume is great if you use it. Fewer, more controlled renders can call for a different pricing model.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The rest of the creative stack.&lt;/strong&gt; Product photos, try-on images, editing an existing video: some teams want those in the same place.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;None of this means one tool is bad. It means the best tool depends on the ad you run next month.&lt;/p&gt;

&lt;h2&gt;
  
  
  The checklist: five things to compare
&lt;/h2&gt;

&lt;p&gt;Open each candidate, use its trial, and write the answers down. Pricing pages and features change often, so date your notes and do not trust old comparison posts, including this one.&lt;/p&gt;

&lt;h3&gt;
  
  
  1. Is your real product in the shot?
&lt;/h3&gt;

&lt;p&gt;Upload your hardest product: glossy packaging, small text, an unusual shape. Generate three scenes. If the label is unreadable or the product changes between clips, the tool is fine for talking heads and wrong for product ads.&lt;/p&gt;

&lt;h3&gt;
  
  
  2. Does the face stay the same?
&lt;/h3&gt;

&lt;p&gt;Save one AI creator and use it in five different scenes. Same face, same hair, same skin tone? If not, you will not build recognition across a campaign.&lt;/p&gt;

&lt;h3&gt;
  
  
  3. What does one usable clip cost?
&lt;/h3&gt;

&lt;p&gt;Not the plan: the clip. Divide what you pay by the clips you would actually run. For reference, on Kreavo a 10-second 720p clip costs $0.92 on Grok Imagine, $2.00 on Wan 3.0, $2.50 on Kling 4.0 and $3.50 on Seedance 2.5 (prices checked 2 November 2026). Do the same maths for every tool you compare.&lt;/p&gt;

&lt;h3&gt;
  
  
  4. Can you choose the video model?
&lt;/h3&gt;

&lt;p&gt;Some tools run one engine for everything. Others let you pick per job. Choice matters for budget (draft on a cheap model, render the winner on a better one) and for quality, because models handle hands, text and lip-sync differently.&lt;/p&gt;

&lt;h3&gt;
  
  
  5. What happens when a render fails?
&lt;/h3&gt;

&lt;p&gt;Every AI video model rejects some prompts or fails some jobs. Ask whether failed renders give the credits back before you scale.&lt;/p&gt;

&lt;h2&gt;
  
  
  A quick way to test any alternative
&lt;/h2&gt;

&lt;p&gt;Give two tools the same 20-minute brief:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;One product photo (the hardest one you sell).&lt;/li&gt;
&lt;li&gt;One 15-second script, three hooks.&lt;/li&gt;
&lt;li&gt;9:16, at the resolution you will actually run.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Score each clip on product fidelity, natural hands and mouth, the first two seconds, and cost per usable clip. Keep the clips. They are also your first ad test.&lt;/p&gt;

&lt;h2&gt;
  
  
  Where Kreavo fits
&lt;/h2&gt;

&lt;p&gt;Kreavo is built for the second kind of ad: one recurring AI influencer showing your real product. You design the creator once, add your product photo, describe the scene, and render it on one of 13 video models priced per clip. Two things it does that are useful if you are coming from a URL-to-ad workflow:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Recreate a format that already works.&lt;/strong&gt; Paste a public TikTok or Instagram Reel and get the same video back (moves, cuts, sound) with your AI influencer and, optionally, your product. It costs $0.29 per second of video, up to 60 seconds.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Edit instead of regenerate.&lt;/strong&gt; Swap the person or the product in a video you already have, billed on the source length (from $0.29 per second at 720p).&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;Video starts with a $1, 7-day trial; images start with 150 free credits. If batch variations from a URL are what you need most, stay where you are. If you want a consistent creator and your real product, our &lt;a href="https://kreavo.ai/creatify-alternative" rel="noopener noreferrer"&gt;Creatify alternative page&lt;/a&gt; walks through the differences, and &lt;a href="https://kreavo.ai/recreate-tiktok" rel="noopener noreferrer"&gt;Recreate a TikTok&lt;/a&gt; is the quickest thing to try.&lt;/p&gt;

&lt;h2&gt;
  
  
  Common mistakes when switching
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Comparing plan prices instead of clip prices.&lt;/strong&gt; A cheaper plan that needs three renders per usable clip is not cheaper.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Testing with an easy product.&lt;/strong&gt; Every tool looks good with a plain box. Test with the product that will break it.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Changing tool and message at the same time.&lt;/strong&gt; Re-run one ad that already worked on the new tool first, so you know what changed.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Ignoring the creator's consistency.&lt;/strong&gt; It does not show in one clip. It shows in clip twenty.&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;If you have tested Creatify against another tool on a physical product, the comments are the most useful part of posts like this. What broke first?&lt;/p&gt;

</description>
      <category>marketing</category>
      <category>saas</category>
      <category>startup</category>
      <category>ai</category>
    </item>
    <item>
      <title>What AI video generation really costs per second in 2026 (13 models, open data)</title>
      <dc:creator>Kreavo</dc:creator>
      <pubDate>Thu, 29 Oct 2026 19:57:11 +0000</pubDate>
      <link>https://dev.to/kreavo/what-ai-video-generation-really-costs-per-second-in-2026-13-models-open-data-ed</link>
      <guid>https://dev.to/kreavo/what-ai-video-generation-really-costs-per-second-in-2026-13-models-open-data-ed</guid>
      <description>&lt;blockquote&gt;
&lt;p&gt;&lt;strong&gt;Short answer:&lt;/strong&gt; at 720p, AI video generation costs between about &lt;strong&gt;$0.06 and $0.63 per second&lt;/strong&gt; of finished video in October 2026, depending on the model. A 10-second 720p clip costs &lt;strong&gt;$0.92&lt;/strong&gt; on Grok Imagine, &lt;strong&gt;$2.00&lt;/strong&gt; on Wan 3.0, &lt;strong&gt;$2.50&lt;/strong&gt; on Kling 4.0, &lt;strong&gt;$3.50&lt;/strong&gt; on Seedance 2.5 and &lt;strong&gt;$3.67&lt;/strong&gt; on Kling 2.6. VEO 3.1 only makes 8-second clips: &lt;strong&gt;$2.00&lt;/strong&gt; (Fast) or &lt;strong&gt;$5.00&lt;/strong&gt; (Quality).&lt;/p&gt;
&lt;/blockquote&gt;

&lt;p&gt;Full disclosure first: we build Kreavo, a platform that runs these models for UGC-style video ads. The numbers below are the prices our own billing formula charges (checked on 29 October 2026), not the model vendors' own API rates. We publish them as an open dataset under CC BY 4.0 because we could not find one place that listed them per second, side by side, with the length and resolution options that actually exist.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why "per second" is the only fair unit
&lt;/h2&gt;

&lt;p&gt;Model pages love "from $X per video". That number hides two things: how long the video is and at what resolution. A $1 clip can be 4 seconds at 480p or 10 seconds at 720p. If you are budgeting a product, a pipeline or an ad test, divide by the seconds you get. Everything below does that.&lt;/p&gt;

&lt;h2&gt;
  
  
  Price per second at 720p
&lt;/h2&gt;

&lt;p&gt;Lowest price per second at 720p for the main versions of each model (for flat-priced versions, the longest clip gives the lowest rate). A few extra variants, such as Grok Imagine 1.5 or Vidu Q3 Turbo, are left out here to keep the table readable; the dataset has all of them and the script further down prints the full list:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Model version&lt;/th&gt;
&lt;th&gt;USD per second&lt;/th&gt;
&lt;th&gt;Lengths offered&lt;/th&gt;
&lt;th&gt;Notes&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Wan 2.6 Turbo&lt;/td&gt;
&lt;td&gt;from $0.056&lt;/td&gt;
&lt;td&gt;5, 10, 15 s&lt;/td&gt;
&lt;td&gt;start image required, flat $0.84 per clip&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Grok Imagine&lt;/td&gt;
&lt;td&gt;$0.092&lt;/td&gt;
&lt;td&gt;6, 10, 15, 30 s&lt;/td&gt;
&lt;td&gt;text-to-video&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;HunyuanVideo 1.5&lt;/td&gt;
&lt;td&gt;$0.134&lt;/td&gt;
&lt;td&gt;5, 8, 10 s&lt;/td&gt;
&lt;td&gt;start image required, silent&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Gemini Omni Flash&lt;/td&gt;
&lt;td&gt;$0.198 to $0.248&lt;/td&gt;
&lt;td&gt;4, 6, 8, 10 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Wan 3.0&lt;/td&gt;
&lt;td&gt;$0.20&lt;/td&gt;
&lt;td&gt;2 to 30 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Seedance 2.0 Fast+&lt;/td&gt;
&lt;td&gt;$0.20&lt;/td&gt;
&lt;td&gt;4 to 15 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Kling 4.0&lt;/td&gt;
&lt;td&gt;$0.25&lt;/td&gt;
&lt;td&gt;3 to 30 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;VEO 3.1 Fast&lt;/td&gt;
&lt;td&gt;$0.25&lt;/td&gt;
&lt;td&gt;8 s only&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Gemini Omni 1.1&lt;/td&gt;
&lt;td&gt;$0.25&lt;/td&gt;
&lt;td&gt;10 s only&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Seedance 2.0+&lt;/td&gt;
&lt;td&gt;$0.27&lt;/td&gt;
&lt;td&gt;4 to 15 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Kling 3.0&lt;/td&gt;
&lt;td&gt;$0.28&lt;/td&gt;
&lt;td&gt;5, 10, 15 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Vidu Q3 Pro&lt;/td&gt;
&lt;td&gt;$0.30&lt;/td&gt;
&lt;td&gt;5, 8, 10, 12, 16 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Wan 3.0 Fast&lt;/td&gt;
&lt;td&gt;$0.32&lt;/td&gt;
&lt;td&gt;2 to 30 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Seedance 2.5&lt;/td&gt;
&lt;td&gt;$0.35&lt;/td&gt;
&lt;td&gt;4 to 30 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Kling 2.6&lt;/td&gt;
&lt;td&gt;$0.367&lt;/td&gt;
&lt;td&gt;5, 10 s&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;VEO 3.1 Quality&lt;/td&gt;
&lt;td&gt;$0.625&lt;/td&gt;
&lt;td&gt;8 s only&lt;/td&gt;
&lt;td&gt;&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Unless noted, these versions generate sound with the video. The spread from the cheapest image-to-video option to the most expensive is about &lt;strong&gt;11x&lt;/strong&gt;. If you need pure text-to-video, the spread is still about &lt;strong&gt;7x&lt;/strong&gt; (Grok Imagine at $0.092 vs VEO 3.1 Quality at $0.625).&lt;/p&gt;

&lt;h2&gt;
  
  
  1080p and 4K
&lt;/h2&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Model version&lt;/th&gt;
&lt;th&gt;USD per second at 1080p&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Wan 2.6 Turbo&lt;/td&gt;
&lt;td&gt;from $0.056 (start image required, same flat $0.84 per clip as at 720p)&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Gemini Omni Flash&lt;/td&gt;
&lt;td&gt;$0.198 to $0.248 (same price as at 720p)&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Vidu Q3 Pro&lt;/td&gt;
&lt;td&gt;$0.35&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Kling 4.0&lt;/td&gt;
&lt;td&gt;$0.37&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Kling 3.0 Pro&lt;/td&gt;
&lt;td&gt;about $0.37&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Gemini Omni 1.1&lt;/td&gt;
&lt;td&gt;$0.38&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Wan 3.0&lt;/td&gt;
&lt;td&gt;$0.40&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Seedance 2.5&lt;/td&gt;
&lt;td&gt;$0.49&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Seedance 2.0+&lt;/td&gt;
&lt;td&gt;$0.60&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Wan 3.0 Fast&lt;/td&gt;
&lt;td&gt;$0.63&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;At 4K, Gemini Omni Flash costs $0.33 to $0.58 per second ($2.31 for 4 s, $3.30 for 10 s) and Gemini Omni 1.1 costs $0.75 per second ($7.50 for its 10-second clip).&lt;/p&gt;

&lt;p&gt;Going from 720p to 1080p adds &lt;strong&gt;40 %&lt;/strong&gt; on Seedance 2.5, &lt;strong&gt;48 %&lt;/strong&gt; on Kling 4.0 and &lt;strong&gt;doubles&lt;/strong&gt; the price on Wan 3.0, while Wan 2.6 Turbo and Gemini Omni Flash charge the same at both. If the video ends up in a 9:16 feed on a phone, test whether your audience can tell before you pay for it.&lt;/p&gt;

&lt;h2&gt;
  
  
  Three things that move the bill more than the model
&lt;/h2&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Resolution.&lt;/strong&gt; See above. It is often a bigger lever than switching models.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Length granularity.&lt;/strong&gt; Fixed-length models (VEO 3.1 at 8 s, Gemini Omni 1.1 at 10 s) make you pay for seconds you may cut. Flat-priced ones (Wan 2.6 Turbo at $0.84 for 5, 10 or 15 s) get cheaper per second the longer you go.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;The start image.&lt;/strong&gt; Several of the cheapest options need a first frame. Generating that image is a separate cost (on our side, $0.07 to $0.75 per image depending on the image model and resolution), so compare the pair, not the clip alone.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;And one that no price table shows: &lt;strong&gt;retries&lt;/strong&gt;. A clip you throw away costs exactly as much as one you keep. Budget per usable clip, not per generation.&lt;/p&gt;

&lt;h2&gt;
  
  
  How long you wait
&lt;/h2&gt;

&lt;p&gt;Render time only counts if enough real jobs back it. Our bar is at least 10 customer jobs from 3 or more customers in the window. In the 30 days to 22 October 2026, two versions passed it (request to finished file, queue included, all lengths and resolutions together):&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Grok Imagine:&lt;/strong&gt; median 2 min 27 s, 90th percentile 5 min 48 s&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Seedance 2.5:&lt;/strong&gt; median 5 min 1 s, 90th percentile 10 min 6 s&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;That is roughly 14 and 22 seconds of waiting per second of video. We do not publish a number for the other versions until they have enough jobs. These figures are the dated 22 October snapshot; the JSON below is refreshed more often, so its render-time fields can show a newer window.&lt;/p&gt;

&lt;h2&gt;
  
  
  Pull the data yourself
&lt;/h2&gt;

&lt;p&gt;Every version in the index carries a &lt;code&gt;price_grid&lt;/code&gt; with every length and resolution it supports. This prints every version available at 720p, sorted, including the variants left out of the table above (Node 18+, no dependencies):&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight javascript"&gt;&lt;code&gt;&lt;span class="c1"&gt;// Cheapest price per second at 720p for every AI video model version in the index&lt;/span&gt;
&lt;span class="kd"&gt;const&lt;/span&gt; &lt;span class="nx"&gt;res&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="k"&gt;await&lt;/span&gt; &lt;span class="nf"&gt;fetch&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="dl"&gt;"&lt;/span&gt;&lt;span class="s2"&gt;https://kreavo.ai/api/model-index.json&lt;/span&gt;&lt;span class="dl"&gt;"&lt;/span&gt;&lt;span class="p"&gt;);&lt;/span&gt;
&lt;span class="kd"&gt;const&lt;/span&gt; &lt;span class="nx"&gt;index&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="k"&gt;await&lt;/span&gt; &lt;span class="nx"&gt;res&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;json&lt;/span&gt;&lt;span class="p"&gt;();&lt;/span&gt;

&lt;span class="kd"&gt;const&lt;/span&gt; &lt;span class="nx"&gt;rows&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="p"&gt;[];&lt;/span&gt;
&lt;span class="k"&gt;for &lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="kd"&gt;const&lt;/span&gt; &lt;span class="nx"&gt;model&lt;/span&gt; &lt;span class="k"&gt;of&lt;/span&gt; &lt;span class="nx"&gt;index&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;models&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;filter&lt;/span&gt;&lt;span class="p"&gt;((&lt;/span&gt;&lt;span class="nx"&gt;m&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="o"&gt;=&amp;gt;&lt;/span&gt; &lt;span class="nx"&gt;m&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;kind&lt;/span&gt; &lt;span class="o"&gt;===&lt;/span&gt; &lt;span class="dl"&gt;"&lt;/span&gt;&lt;span class="s2"&gt;video&lt;/span&gt;&lt;span class="dl"&gt;"&lt;/span&gt;&lt;span class="p"&gt;))&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
  &lt;span class="k"&gt;for &lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="kd"&gt;const&lt;/span&gt; &lt;span class="nx"&gt;v&lt;/span&gt; &lt;span class="k"&gt;of&lt;/span&gt; &lt;span class="nx"&gt;model&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;versions&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;
    &lt;span class="kd"&gt;const&lt;/span&gt; &lt;span class="nx"&gt;at720&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="nx"&gt;v&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;price_grid&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;rows&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;filter&lt;/span&gt;&lt;span class="p"&gt;(([,&lt;/span&gt; &lt;span class="nx"&gt;resolution&lt;/span&gt;&lt;span class="p"&gt;])&lt;/span&gt; &lt;span class="o"&gt;=&amp;gt;&lt;/span&gt; &lt;span class="nx"&gt;resolution&lt;/span&gt; &lt;span class="o"&gt;===&lt;/span&gt; &lt;span class="dl"&gt;"&lt;/span&gt;&lt;span class="s2"&gt;720p&lt;/span&gt;&lt;span class="dl"&gt;"&lt;/span&gt;&lt;span class="p"&gt;);&lt;/span&gt;
    &lt;span class="k"&gt;if &lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="o"&gt;!&lt;/span&gt;&lt;span class="nx"&gt;at720&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;length&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="k"&gt;continue&lt;/span&gt;&lt;span class="p"&gt;;&lt;/span&gt;
    &lt;span class="kd"&gt;const&lt;/span&gt; &lt;span class="nx"&gt;perSecond&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="nb"&gt;Math&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;min&lt;/span&gt;&lt;span class="p"&gt;(...&lt;/span&gt;&lt;span class="nx"&gt;at720&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;map&lt;/span&gt;&lt;span class="p"&gt;(([&lt;/span&gt;&lt;span class="nx"&gt;seconds&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="nx"&gt;usd&lt;/span&gt;&lt;span class="p"&gt;])&lt;/span&gt; &lt;span class="o"&gt;=&amp;gt;&lt;/span&gt; &lt;span class="nx"&gt;usd&lt;/span&gt; &lt;span class="o"&gt;/&lt;/span&gt; &lt;span class="nx"&gt;seconds&lt;/span&gt;&lt;span class="p"&gt;));&lt;/span&gt;
    &lt;span class="nx"&gt;rows&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;push&lt;/span&gt;&lt;span class="p"&gt;({&lt;/span&gt; &lt;span class="na"&gt;version&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="nx"&gt;v&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;name&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="na"&gt;usd_per_s&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="o"&gt;+&lt;/span&gt;&lt;span class="nx"&gt;perSecond&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;toFixed&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="mi"&gt;3&lt;/span&gt;&lt;span class="p"&gt;),&lt;/span&gt; &lt;span class="na"&gt;start_image&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="nx"&gt;v&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;requires_image&lt;/span&gt; &lt;span class="p"&gt;});&lt;/span&gt;
  &lt;span class="p"&gt;}&lt;/span&gt;
&lt;span class="p"&gt;}&lt;/span&gt;
&lt;span class="nx"&gt;rows&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;sort&lt;/span&gt;&lt;span class="p"&gt;((&lt;/span&gt;&lt;span class="nx"&gt;a&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="nx"&gt;b&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="o"&gt;=&amp;gt;&lt;/span&gt; &lt;span class="nx"&gt;a&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;usd_per_s&lt;/span&gt; &lt;span class="o"&gt;-&lt;/span&gt; &lt;span class="nx"&gt;b&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nx"&gt;usd_per_s&lt;/span&gt;&lt;span class="p"&gt;);&lt;/span&gt;
&lt;span class="nx"&gt;console&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;table&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="nx"&gt;rows&lt;/span&gt;&lt;span class="p"&gt;);&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Swap &lt;code&gt;"720p"&lt;/code&gt; for &lt;code&gt;"1080p"&lt;/code&gt; or &lt;code&gt;"480p"&lt;/code&gt; to get the other tables. The data is CC BY 4.0: reuse it, just credit the source.&lt;/p&gt;

&lt;h2&gt;
  
  
  A monthly budget example
&lt;/h2&gt;

&lt;p&gt;Say you need 100 ten-second 720p clips a month:&lt;/p&gt;

&lt;div class="table-wrapper-paragraph"&gt;&lt;table&gt;
&lt;thead&gt;
&lt;tr&gt;
&lt;th&gt;Model&lt;/th&gt;
&lt;th&gt;Monthly cost&lt;/th&gt;
&lt;/tr&gt;
&lt;/thead&gt;
&lt;tbody&gt;
&lt;tr&gt;
&lt;td&gt;Grok Imagine&lt;/td&gt;
&lt;td&gt;$92&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Wan 3.0&lt;/td&gt;
&lt;td&gt;$200&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Kling 4.0&lt;/td&gt;
&lt;td&gt;$250&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;Seedance 2.5&lt;/td&gt;
&lt;td&gt;$350&lt;/td&gt;
&lt;/tr&gt;
&lt;tr&gt;
&lt;td&gt;VEO 3.1 Quality (100 x 8 s)&lt;/td&gt;
&lt;td&gt;$500&lt;/td&gt;
&lt;/tr&gt;
&lt;/tbody&gt;
&lt;/table&gt;&lt;/div&gt;

&lt;p&gt;Same volume, same resolution, a 5x difference. Most teams end up mixing: a cheap model to find the hook, a more expensive one to render the winner.&lt;/p&gt;

&lt;h2&gt;
  
  
  Caveats, so you can trust the rest
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Dollar figures use 1 credit = $0.01, the rate of the smallest credit pack ($10 for 1,000 credits). Monthly plans and larger packs cost the same or less per credit.&lt;/li&gt;
&lt;li&gt;For a few models (Kling 3.0, VEO 3.1, Hailuo 2.3) the published resolution is the one we measured on the output, not the setting, because the setting is not always honored.&lt;/li&gt;
&lt;li&gt;Prices change. Every row has a snapshot date. In the 30-day window above, every failed job that had been charged was refunded.&lt;/li&gt;
&lt;li&gt;Price says nothing about quality. Run the same prompt on two or three models before you commit a budget.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Where Kreavo fits
&lt;/h2&gt;

&lt;p&gt;Kreavo is an AI UGC video tool for ecommerce brands: you keep one AI influencer, add your product photo, and render the scene on whichever of these models fits the budget. The full table, the per-second grid and two CSV files live on the &lt;a href="https://kreavo.ai/ai-video-generation-cost-index" rel="noopener noreferrer"&gt;AI video generation cost index&lt;/a&gt;, with the snapshot date on every row.&lt;/p&gt;

&lt;p&gt;If you spot a number that looks wrong, leave a comment with the model, length and resolution, and we will check it against the formula.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>video</category>
      <category>api</category>
      <category>datascience</category>
    </item>
  </channel>
</rss>
