Load Light

What it does

Drop in stills, video, or a mixed batch. Load Light identifies the media, lets you choose the output format and quality, and builds the responsive sizes the page needs. You can start with a preset or tune the settings yourself, then export a ZIP or download the files individually.

The export can also include an HTML media reference with filenames, dimensions, metadata, and output order. That saves a second round of digging when it is time to wire the assets into a page.

Why encode in the browser

One source file may need to become a thumbnail, a hero asset, and several responsive versions. Keeping that work in a browser tab makes it easy to try a setting, check the estimated savings, and adjust it again while the design is still in front of you.

The files stay on your machine while Load Light processes them. Large video jobs still depend on the available CPU, GPU, and memory, but image batches are usually quick to turn around.

AI Max gives an LLM the media context

AI Max exports one token-optimized JSON file for the complete batch. It maps each asset to a consistent schema, groups responsive sizes in nested arrays, and includes dominant color palettes.

The regular HTML reference is useful when you want markup. AI Max is for handing the batch to an LLM or coding assistant. The model can read one compact media map instead of searching the asset folders, matching variants by filename, and piecing the same context together across several files. AI Max documents the media; it does not change the compression settings or generate another encode.

How I use it in a workflow

• Start with a size target, then bring the quality up until the file holds together at the dimensions it will actually use.
• Name and group outputs by target (hero, inline, fallback) so downstream pages stay consistent.
• Keep the grade and finishing work upstream. Load Light handles bytes, sizes, and delivery.
• Export AI Max when an LLM will help build or document the page, especially when the batch has many assets and responsive variants.

Technical notes

Load Light uses the browser's current decoding and encoding capabilities, including the WebCodecs GPU path when it is available. The UI calls out formats or processing paths that the browser cannot use. File handling stays local to the page unless you later connect the tool to your own pipeline.