AVOCADO (Adaptive Video Optimisation: Crawler-Accessible Deterministic Output)
Let AI search engines discover your interactive video, while keeping the audience, the engagement, and the data on your own site.
What AVOCADO is and what it does
AVOCADO turns your interactive video into structured data that AI systems can read.
The problem it solves
AI search and answer engines (Google AI Overview, ChatGPT, Perplexity and others) are changing how people find information. For text, this has been disruptive: an engine scrapes a page, paraphrases it, and answers the question directly. The user gets their answer, and the publisher loses the click, the attention, and often the visitor entirely.
What it actually does
Every project built in Stornaway can be indexed by AI crawlers using a standard JSON-LD schema, the same kind of structured data the web already uses to help search engines understand pages and articles, applied here to the structure and content of interactive video.
This means an AI engine can understand what your project contains (its structure, branches, chapters and interactive moments), match a viewer’s question to the exact moment that answers it, and return a deep link into that moment rather than just a link to the video’s homepage.
How to set it up
You need to add two things to a page on your own website: the Stornaway player, and the AVOCADO code that indexes it.
Step 1
Build your project in Stornaway. Create your interactive video. Make sure each section has captions and each island has a relevant name and description, as this is what AVOCADO uses to describe each moment to the AI.
When you upload your video files, don’t burn the subtitles into the video itself. Instead, include your captions as separate .vtt files alongside each clip. You can read more about how to upload your captions here.
Step 2
Add the player to a page on your site.
Copy the embed code from Stornaway and paste it into a page you own, so the video lives inside your branding and context.
(If your video isn’t being embedded anywhere and just lives on its Stornaway player page at player.stornaway.io/watch/[ID], skip this step.)
Step 3
Drop the AVOCADO code next to it. The AVOCADO (JSON-LD) code is automatically generated for each project and gives the AI a readable index of the project: every section, what’s said in it, and a link that jumps straight to it.
To get it, go to Publish > Share & Export > AVOCADO tab.
Caption languages:
- Select “Select primary language” if you only have one language set up in your captions
- Select “Multiple languages” if you have more than one
Then, which box you tick depends on where your video is shown:
- If the player is embedded on your own web page, tick: “Generate AVOCADO for the web page where you are embedding the player.” This generates code you’ll need to paste onto that page, alongside the player, so AI sends people to your branded page rather than the raw player link.
- If the video isn’t embedded anywhere and just lives on its Stornaway player page, tic: “Show AVOCADO on player page” instead.
Step 4
Publish, and let AI find it. Once live, index your page on Google Search Console, then wait until Google has indexed it. After that, AI engines can crawl it, match questions to moments, and build deep links to your page.
See it in action
Why it matters
AI answer engines are changing how people find things. AVOCADO makes sure that shift works for you, not against you.
Keep your audience. Viewers land on your page, not YouTube or a Google answer box. You keep the traffic, the branding and the relationship.
Can’t be scraped away. Unlike text, an interactive video can’t be summarised into a standalone AI answer. The engine has to send people to you to watch it.
Deep links to the exact moment. AI matches a viewer’s question to the precise point in your video that answers it, and drops them straight there.
Engagement you can measure. Because it all happens on your platform, you can track, optimise and grow the engagement instead of handing it to someone else’s algorithm.
Ready for your AI agents. Your help bots and learning assistants can answer with the actual moment in a training video, not just a link to the whole thing.