A saved description, tags and subject position for an image in your workspace.
About
Looks at an image in your workspace (PNG, JPEG, WebP or GIF) and records what it shows: a description of up to 12 words, three to six tags, a line on when to use it, and where the subject sits in the frame. The subject position is what lets a layout crop to the face or the product instead of the centre.
It is for images that arrived without labels, or whose labels are out of date. The result is saved as metadata beside the image, which is what search and picture-choosing tools read.
The reading is a paid call to a vision model. When the caller can look at images itself, it can pass its own reading instead and no model call is made. It currently runs inside Syvon and needs a Syvon workspace.
- TypeScript
- AI vision models
Requirements
- Runs with
- Node.js 20 or later
- Operating systems
- windows, macos, linux
- Network
- Sends the image to the vision provider over the internet.
Costs
- External costs
- OpenAI: billed to your own key — Planned direct lane: you would bring your own OpenAI key. Today the call runs through the Syvon provider proxy and Syvon credits.