One of the most common issues users report is the AI describing media inaccurately — for example, saying "check out this video of me in the shower" when the attached media is actually a photo taken outdoors. This isn't a bug. It's a media context problem.
The AI cannot see your photos or videos directly. It relies on a text field called Context attached to each media file to understand what it contains. When that context is missing, vague, or wrong, the AI fills in the blanks — and gets it wrong.
Every media file in your Vault has a Context field — a short text description (up to 300 characters) that tells the AI what the media looks like. When the AI sends a message template to a fan, it receives the media type, duration (for videos), and context. If the Context field is empty, the AI has zero information about what the fan is about to see.
Open the Vault, select any media, and you'll see the Context field in the edit panel on the right. Type a short description of what the media shows.
When editing a message template in the Script editor, click on any attached media. The Context field appears below the media preview.
Be specific and visual. Describe what someone would actually see: setting, outfit, pose, activity.
Keep it under 300 characters. Focus on the most important visual details.
Don't write what the AI should say. Write what the media contains.
Avoid vague terms like "Sexy photo" — "Red lingerie on bed, looking over shoulder" tells the AI everything.
