docs: document upload and that embedding atlas supports embedding and projection (#13)

Co-authored-by: Fred Hohman <[email protected]>

Author: domoritz

packages/docs/algorithms.md

@@ -1,6 +1,6 @@
 # Algorithms
 
-The `embedding-atlas` package contains some useful algorithms regarding embeddings and clustering.
+The `embedding-atlas` package contains some useful algorithms for computing embeddings and clustering.
 
 ## UMAP
 

packages/docs/overview.md

@@ -2,16 +2,18 @@
 
 Embedding Atlas is a tool that provides interactive visualizations for large embeddings. It allows you to visualize, cross-filter, and search embeddings and metadata.
 
+You can use Embedding Atlas directly from this website by [loading your own data](https://apple.github.io/embedding-atlas/upload/). Embedding Atlas will compute the embedding and projection in your browser. Your data does not leave your machine.
+
 Embedding Atlas is released as two packages:
 
 - A Python package `embedding-atlas` that provides:
-
   - A [command line utility](./tool.md) for running Embedding Atlas on a data frame.
   - A [Jupyter widget](./widget.md) for using Embedding Atlas in Jupyter notebooks.
   - A [Streamlit component](./streamlit.md) for using Embedding Atlas in Streamlit apps.
 
-- An npm package `embedding-atlas` that exposes the user interface components as API so you can use them in your own applications. Below are the exposed components:
+All of these approaches allow you to compute embeddings (with custom models) and projections.
 
+- An npm package `embedding-atlas` that exposes the user interface components as API so you can use them in your own applications. Below are the exposed components:
   - [Table](./table.md)
   - [EmbeddingView](./embedding-view.md)
   - [EmbeddingViewMosaic](./embedding-view-mosaic.md)