Sloth Notes: a local AI digest for the notes you don't want to reread

Description and use

Sloth Notes is a local web app that condenses meeting notes by extracting source‑linked highlights and action suggestions, reducing text length by 54 % and running entirely in the browser with a MiniLM embedding model.

Problem

Rereading meeting recaps to find next actions is busywork and inefficient.

Outcome

On a synthetic study‑group recap of 127 words, Sloth Notes reduced the text to 59 words (54% reduction) in ~0.5 s after model assets were cached, and surfaced four keyword‑based action suggestions.

Workflow

• Paste recap into the web interface
• Select number of highlights (3,5, or 7)
• Click "Make it shorter"
• The app downloads the quantized MiniLM model on first run
• Splits text into sentences with Intl.Segmenter
• Generates sentence embeddings in a Web Worker
• Computes document centroid
• Scores sentences balancing relevance and diversity
• Selects top sentences and restores original order
• Displays source reference buttons for each highlight
• Shows keyword‑based action suggestions
• Exports digest as Markdown

Evidence

Sloth Notes turns a messy block of English notes into a short, source-linked digest and a list of suggested actions.

Source

The first run downloads the quantized model and runtime files. The application then computes sentence embeddings on your device.

Source

In the browser check, the synthetic study-group recap went from 127 words to 59 words, a 54% reduction, in about 0.5 seconds with model assets already cached.

Source

The model supplies the meaning-based vectors that drive selection. The follow-up list is a separate, explicitly labelled keyword heuristic.

Source

The app accepts up to 50,000 characters and 400 sentences, rejects oversized input without silently cutting it, and exports a Markdown digest.

Source