Skip to content

What this extension adds

Stockly reads supplier PDFs and guides pickers by voice on its own. This extension gives both jobs a language model: a document can be handed to the model instead of being described column by column, and the packing assistant can speak in a generated voice instead of the browser’s own.

Nothing here replaces a part of Stockly — every screen below is an addition that appears once the extension is installed.

The extension has no menu of its own. Its pages sit inside Stockly’s menu, next to the ones they belong with:

ScreenWhere
Connections to a providerStockly → Settings → AI Assistant
Prompt templatesStockly → Settings → Templates → AI templates
Generated speechStockly → Settings → Diagnostics → Voice clips
Record of every model callStockly → Settings → Diagnostics → AI API logs

Inside a document, the extension shows up as Analyze with AI and as the AI (OpenAI) engine.

Every call is billed by the provider, and what is analysed — the document, its pages, the phrases to speak — is sent to them.

That is worth stating plainly next to a common misreading: choosing the classic reader instead does not keep the document at home. Stockly’s own reader sends it to Adobe PDF Services and is billed there. Reading a PDF without any external service is not something either route offers, so the question is which provider, not whether.

Import in Stockly

Manage AI connections and API keys covers the connections, the templates and the clip library.

View and clear the AI call log (contains prompts and model responses) is separate on purpose: those rows carry the prompt, the document text and the model’s answer, which is a different level of exposure than the fact that a call happened.

Permissions in Stockly