MAISIGHT REV 01

BLUEPRINT · 01. EXTENSION

Drop mAiSight onto
any technical page.

Install the browser extension, open any docs page, blog post, or arXiv mirror, and mAiSight pulls the glossary out and wraps every defined term in the prose — hover a term for its definition, hover a rail row to light up every place it appears.

Download the extension maisight-extension.zip · v1.2 · ~120 KB

Manifest V3 · Chromium-based browsers. Works on docs sites, blog posts, arXiv mirrors — anything text-shaped.

How to load it  
  1. Unzip maisight-extension.zip.
  2. Open chrome://extensions and toggle Developer mode.
  3. Click Load unpacked and pick the unzipped folder.
  4. Visit any technical article and click the mAiSight toolbar icon.

VIEW MODE

The dominant sequence transduction models are based on complex recurrent or convolutional neural networks that include an encoder and a decoder. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely.

We call our particular attention “Scaled Dot-Product Attention”. The input consists of queries and keys of dimension dk, and values of dimension dv. We compute the dot products of the query with all keys, divide each by √dk, and apply a softmax function to obtain the weights on the values.

Instead of performing a single attention function with dmodel-dimensional keys, values and queries, we found it beneficial to linearly project the queries, keys and values h times with different, learned linear projections. Multi-head attention allows the model to jointly attend to information from different representation subspaces at different positions.

Self-attention, sometimes called intra-attention, is an attention mechanism relating different positions of a single sequence in order to compute a representation of the sequence. Since the model contains no recurrence and no convolution, to make use of the order of the sequence we add a positional encoding to the input embeddings at the bottoms of the encoder and decoder stacks.

Hover a marked term — or any rail row — to see the binding. Excerpt · Vaswani et al., 2017 · arXiv:1706.03762

BLUEPRINT · 02. FOR DOCS TEAMS

Simple as 1 · 2 · 3.

Shipping API, SDK, or product docs? Point mAiSight at your sources and it becomes the reading surface — published as a hosted page or an embed widget that drops onto your existing docs site.

  1. 01

    Ingest

    Drop in Markdown, an OpenAPI spec, an arXiv bundle, or any web page. mAiSight parses it into structured JSON.

  2. 02

    Extract

    Claude proposes a glossary — accept, edit, or reject each term before publish. Every row shows its origin: spec, AI, or edit.

  3. 03

    Read

    Publish as a hosted page, a same-origin embed, or a browser overlay. Every defined term lights up on both sides.