A large language model (LLM) friendly embed code is designed to make your videos more accessible to LLMs, increasing the chances that your videos will appear in AI search results. This embed code works with both video and audio embeds.
The LLM-friendly embed code appears the same way an inline embed code does on a webpage—the media plays inline, with the rest of your content.
Tip
LLM embed codes currently work best with ChatGPT and Perplexity.
LLM-friendly embed codes require a transcript, as the transcript is injected directly into the embed code for LLMs to read and surface in search results. Without a transcript, the LLM option will not appear.
Note
The media's transcript is included in LLM embed codes. If you update the transcript, the embed code will not dynamically update to match the new transcript. If you make significant updates to the transcript (including media replacements), swap out the old embed code for a new one so LLMs can read and surface the correct content.
Get started with LLM-friendly embeds
The Embed modal includes an "LLM-Friendly" tab where you can generate an LLM embed code. This is a standard, inline embed code that can be responsive or fixed size.
Captions are turned on by default for LLM-friendly embeds, even if your player's captions are set to "off by default." This keeps the transcript that loads before playback consistent with the captions shown during playback, which helps protect your page's SEO. You can remove the captions control from the player completely, but we don't recommend it for this embed type.
Tip
If your media has multiple transcript languages, the LLM-friendly embed includes the transcript in the media's default language.
If you'd like to see a preview of your embed code before adding it to your webpage, click "Show Embed Code." When you're ready, click "Copy Code" and paste it into your website builder.
Note
LLM-friendly embed codes are not compatible with iframe or popover embeds, webinars, or legacy (E-v1) embed codes. Chapters are not included in the embedded transcript.
FAQ and troubleshooting
Do LLM embed codes include SEO metadata?
LLM-friendly embed codes include SEO metadata by default and cannot be turned off.
Can I use an LLM embed code for my Channel or playlist?
LLM-friendly embed codes are not compatible with Channels or playlists and currently work only with video and audio files.
Will my existing embeds automatically become LLM-friendly?
No, existing embed codes do not update automatically. To make a video LLM-friendly, copy the LLM-friendly embed code from the Embed menu and replace the old embed code on your page. To switch back, replace it with the Inline embed code.
How do I confirm that my LLM embed is working?
Open your published page, right-click, and choose View Page Source (not "Inspect"). Search for a phrase you know is in the video. If you find it, the transcript is on the page for AI crawlers to read. Browser-based assistants like Copilot read the page after the Wistia player loads and replaces the transcript with the video, so they aren't a reliable test. Make sure the page is public, since AI crawlers can't reach pages behind a login or on staging sites that block crawlers. Then try asking ChatGPT or Perplexity about something only mentioned in the video. New pages can take time to be picked up.
Captions show on my video even though I turned them off
This is expected for LLM-friendly embeds. Captions are turned on so they match the transcript included in the embed code. If you need captions off for a specific video (for example, if your video already has burned-in captions), use the Inline embed code instead.
Do LLM-friendly embed codes affect page speed?
Wistia LLM embeds involve a small tradeoff in page and site speed because they load more content (the transcript). However, since this content is simple (plain HTML text), it won't markedly affect load times.