<?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: Xiao Lin</title>
    <description>The latest articles on DEV Community by Xiao Lin (@xiao_lin_d3a217fb49d23f1b).</description>
    <link>https://dev.to/xiao_lin_d3a217fb49d23f1b</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%2F4257025%2Fa62558c1-8f79-43bb-8465-2167e024a047.png</url>
      <title>DEV Community: Xiao Lin</title>
      <link>https://dev.to/xiao_lin_d3a217fb49d23f1b</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/xiao_lin_d3a217fb49d23f1b"/>
    <language>en</language>
    <item>
      <title>One API Key, Many Routes: Calling AI Video Generation Models in Python (with Automatic Failover)</title>
      <dc:creator>Xiao Lin</dc:creator>
      <pubDate>Sat, 21 Nov 2026 09:29:45 +0000</pubDate>
      <link>https://dev.to/xiao_lin_d3a217fb49d23f1b/one-api-key-many-routes-calling-ai-video-generation-models-in-python-with-automatic-failover-1ncl</link>
      <guid>https://dev.to/xiao_lin_d3a217fb49d23f1b/one-api-key-many-routes-calling-ai-video-generation-models-in-python-with-automatic-failover-1ncl</guid>
      <description>&lt;p&gt;If you're building any kind of content pipeline in 2026 — product videos, social clips, demo reels — &lt;strong&gt;AI video generation&lt;/strong&gt; is now the cheapest way to get usable footage. But wiring it into your backend has a few gotchas that most tutorials skip: model routing, idempotency, and what happens when a generation fails halfway.&lt;/p&gt;

&lt;p&gt;In this post I'll show the exact Python flow I use, against a &lt;strong&gt;multi-route video generation API&lt;/strong&gt; — one key, dozens of models, and multiple independent routes per model.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why "multi-route" matters for video generation
&lt;/h2&gt;

&lt;p&gt;Most API providers give you one endpoint per model. AIShow takes a different approach: &lt;strong&gt;the same video model can be served by several independent routes&lt;/strong&gt; (different upstream providers), and every route is addressable by its own model code.&lt;/p&gt;

&lt;p&gt;That gives you two superpowers as a backend developer:&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;
&lt;strong&gt;Price arbitrage&lt;/strong&gt; — identical model, different routes, different per-second pricing. Pick the cheapest route that supports your resolution.&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Automatic failover&lt;/strong&gt; — if one route is down or times out, the platform switches to the next route for the same logical model. You don't re-submit or re-pay, and one "stuck" provider never poisons your queue.&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;In practice you rarely pick routes manually: you submit to a model, and the platform routes it. But every route has its own visible model code, which is great for cost dashboards.&lt;/p&gt;

&lt;h2&gt;
  
  
  The three-call pattern
&lt;/h2&gt;

&lt;p&gt;All async video generation APIs follow the same shape: &lt;strong&gt;submit → poll → fetch&lt;/strong&gt;.&lt;/p&gt;

&lt;h3&gt;
  
  
  1. List available models
&lt;/h3&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;curl -s https://www.aishow.uno/api/v1/models \
  -H &lt;span class="s2"&gt;"Authorization: Bearer YOUR_API_KEY"&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;You get back every model with its &lt;code&gt;kind&lt;/code&gt; (&lt;code&gt;video&lt;/code&gt; / &lt;code&gt;image&lt;/code&gt; / &lt;code&gt;music&lt;/code&gt; / &lt;code&gt;chat&lt;/code&gt;), price tiers, and supported durations. Video model codes look like &lt;code&gt;aishow-&amp;lt;model&amp;gt;-&amp;lt;resolution&amp;gt;-&amp;lt;id&amp;gt;&lt;/code&gt; — one code per route/resolution combo, so you always know exactly what you're paying for.&lt;/p&gt;

&lt;h3&gt;
  
  
  2. Submit a video task
&lt;/h3&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="k"&gt;import&lt;/span&gt; &lt;span class="n"&gt;requests&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;uuid&lt;/span&gt;

&lt;span class="no"&gt;BASE&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="s2"&gt;"https://www.aishow.uno/api/v1"&lt;/span&gt;
&lt;span class="no"&gt;HDRS&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="s2"&gt;"Authorization"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"Bearer YOUR_API_KEY"&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;

&lt;span class="k"&gt;def&lt;/span&gt; &lt;span class="nf"&gt;submit_video&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;model_code&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="kt"&gt;str&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;prompt&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="kt"&gt;str&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;seconds&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="kt"&gt;int&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="mi"&gt;5&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
    &lt;span class="n"&gt;r&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;requests&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;post&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;f&lt;/span&gt;&lt;span class="s2"&gt;"{BASE}/generate"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;headers&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="no"&gt;HDRS&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;json&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;
        &lt;span class="s2"&gt;"kind"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="s2"&gt;"video"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
        &lt;span class="s2"&gt;"model"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="n"&gt;model_code&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
        &lt;span class="s2"&gt;"prompt"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="n"&gt;prompt&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
        &lt;span class="s2"&gt;"params"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="s2"&gt;"duration"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="n"&gt;seconds&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
        &lt;span class="s2"&gt;"idempotency_key"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="kt"&gt;str&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;uuid&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;uuid4&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="p"&gt;,&lt;/span&gt;
    &lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;timeout&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;30&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
    &lt;span class="n"&gt;r&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;raise_for_status&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
    &lt;span class="k"&gt;return&lt;/span&gt; &lt;span class="n"&gt;r&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="p"&gt;)&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"task_id"&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;The &lt;code&gt;idempotency_key&lt;/code&gt; is essential: network retries with the same key never double-charge you.&lt;/p&gt;

&lt;h3&gt;
  
  
  3. Poll for a terminal state
&lt;/h3&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight python"&gt;&lt;code&gt;&lt;span class="k"&gt;import&lt;/span&gt; &lt;span class="n"&gt;time&lt;/span&gt;

&lt;span class="k"&gt;def&lt;/span&gt; &lt;span class="nf"&gt;wait_video&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;task_id&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="kt"&gt;str&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;timeout&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="kt"&gt;int&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="mi"&gt;900&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
    &lt;span class="n"&gt;deadline&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;time&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;time&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="o"&gt;+&lt;/span&gt; &lt;span class="n"&gt;timeout&lt;/span&gt;
    &lt;span class="k"&gt;while&lt;/span&gt; &lt;span class="n"&gt;time&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;time&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt; &lt;span class="o"&gt;&amp;lt;&lt;/span&gt; &lt;span class="n"&gt;deadline&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
        &lt;span class="n"&gt;t&lt;/span&gt; &lt;span class="o"&gt;=&lt;/span&gt; &lt;span class="n"&gt;requests&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;get&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;f&lt;/span&gt;&lt;span class="s2"&gt;"{BASE}/task"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;headers&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="no"&gt;HDRS&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt;
                         &lt;span class="n"&gt;params&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="p"&gt;{&lt;/span&gt;&lt;span class="s2"&gt;"id"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt; &lt;span class="n"&gt;task_id&lt;/span&gt;&lt;span class="p"&gt;}&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="n"&gt;timeout&lt;/span&gt;&lt;span class="o"&gt;=&lt;/span&gt;&lt;span class="mi"&gt;15&lt;/span&gt;&lt;span class="p"&gt;)&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="p"&gt;)&lt;/span&gt;
        &lt;span class="k"&gt;if&lt;/span&gt; &lt;span class="n"&gt;t&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"status"&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt; &lt;span class="o"&gt;==&lt;/span&gt; &lt;span class="s2"&gt;"success"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
            &lt;span class="k"&gt;return&lt;/span&gt; &lt;span class="n"&gt;t&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"result_url"&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt;
        &lt;span class="k"&gt;if&lt;/span&gt; &lt;span class="n"&gt;t&lt;/span&gt;&lt;span class="p"&gt;[&lt;/span&gt;&lt;span class="s2"&gt;"status"&lt;/span&gt;&lt;span class="p"&gt;]&lt;/span&gt; &lt;span class="o"&gt;==&lt;/span&gt; &lt;span class="s2"&gt;"failed"&lt;/span&gt;&lt;span class="p"&gt;:&lt;/span&gt;
            &lt;span class="k"&gt;raise&lt;/span&gt; &lt;span class="nc"&gt;RuntimeError&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="n"&gt;t&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;get&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="s2"&gt;"error"&lt;/span&gt;&lt;span class="p"&gt;,&lt;/span&gt; &lt;span class="s2"&gt;"generation failed"&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
        &lt;span class="n"&gt;time&lt;/span&gt;&lt;span class="p"&gt;.&lt;/span&gt;&lt;span class="nf"&gt;sleep&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="mi"&gt;10&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
    &lt;span class="k"&gt;raise&lt;/span&gt; &lt;span class="nc"&gt;TimeoutError&lt;/span&gt;&lt;span class="p"&gt;(&lt;/span&gt;&lt;span class="s2"&gt;"video task timed out"&lt;/span&gt;&lt;span class="p"&gt;)&lt;/span&gt;
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Only treat &lt;code&gt;success&lt;/code&gt; / &lt;code&gt;failed&lt;/code&gt; as terminal — everything else means "keep waiting".&lt;/p&gt;

&lt;h2&gt;
  
  
  The draft-first trick that cuts costs 5-10x
&lt;/h2&gt;

&lt;p&gt;Video generation pricing scales with duration and resolution. So don't render 15 seconds of 1080p just to "see if the idea works":&lt;/p&gt;

&lt;ol&gt;
&lt;li&gt;Submit a &lt;strong&gt;5-second draft&lt;/strong&gt; on a cheap route → check composition &amp;amp; motion&lt;/li&gt;
&lt;li&gt;Re-submit the same prompt on an HD route for the final render&lt;/li&gt;
&lt;/ol&gt;

&lt;p&gt;Since each route/resolution is its own model code, switching is literally changing one string.&lt;/p&gt;

&lt;h2&gt;
  
  
  What happens on failure
&lt;/h2&gt;

&lt;p&gt;Generation does fail sometimes (complex prompts, busy routes). Make sure your provider &lt;strong&gt;auto-refunds failed tasks&lt;/strong&gt; — the platform I use credits the points back automatically, so a retry is free to just re-submit. If your provider doesn't do this, build compensation logic into your pipeline.&lt;/p&gt;

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

&lt;ul&gt;
&lt;li&gt;Async video generation = submit → poll → fetch. Template it once, reuse for image/music by changing &lt;code&gt;kind&lt;/code&gt;
&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Multi-route platforms&lt;/strong&gt; give you per-route pricing and automatic failover without extra code&lt;/li&gt;
&lt;li&gt;Always use: idempotency keys, terminal-state polling, refund-aware retries&lt;/li&gt;
&lt;li&gt;Draft-first rendering is the single biggest cost lever&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;The API I used here: &lt;a href="https://www.aishow.uno" rel="noopener noreferrer"&gt;AIShow&lt;/a&gt; — Markdown docs at &lt;a href="https://www.aishow.uno/openapi_md.php?raw=1" rel="noopener noreferrer"&gt;openapi_md.php?raw=1&lt;/a&gt;, LLM-friendly if you want your agent to wire it up for you.&lt;/p&gt;

&lt;p&gt;Happy to answer questions in the comments.&lt;/p&gt;

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