Showcase

On all websites, when hovering over an image, show a preview button in the bottom-right corner of the image. Clicking it opens an image viewer modal that supports:

1. Zooming in and out of the image.
2. Dragging/panning the image.
3. Downloading the image.

If the image has an ancestor link (the image is wrapped in an <a> element), add an "Open link" button next to the download button that opens that link in a new tab.

A multilingual quick-reply plugin.

1. A floating button sits in the bottom-right corner; click it to open a popup where you can enter the source content and the reply topic — fill in either one or both.
2. After confirming, it calls the LLM and generates a reply using the configured LLM API in the target language for the target content.
3. The reply is split into a target-language section and a local-language section, which is the translation in the target language, not a summary, and the target-language section can be copied with one click.
4. At the bottom of the popup is a skill list selector (single-select). If a skill is selected, its prompt is added to the context, with a Settings button beside it — click to edit or create new skills, and the edited/created skills need to be saved.
5. After the reply is generated, you can enter adjustment feedback to regenerate it.
LLM configuration: baseURL: <your-base-url>, API key: <your-api-key>, and model: <your-model>.

Use https://github.com/sam70361/aora-bot to build a page pet. Clicking it should trigger interactions, and its appearance should be customizable.

Embed www.youtube.com