Embedding Atlas
Open-source toolkit from Apple for interactive visualization and exploration of large embeddings in the browser, Python (CLI/Jupyter), and Streamlit.
At a Glance
Get started with Embedding Atlas at no cost
Engagement
Available On
Updated Feb 2026
About Embedding Atlas
Embedding Atlas is an MIT-licensed toolkit for working with large text or image embeddings. It provides interactive, in-browser visualizations with cross-filtering and metadata search, and ships as both a Python package (CLI tool, Jupyter widget, Streamlit component) and an npm UI component library. It computes embeddings and 2D projections locally (data stays on-device), and includes WebAssembly implementations of UMAP, approximate nearest-neighbor search, and density-based clustering. Integrations include SentenceTransformers for embedding generation and optional loading of datasets from Hugging Face.
Community Discussions
Be the first to start a conversation about Embedding Atlas
Share your experience with Embedding Atlas, ask questions, or help others learn from your insights.
Pricing
Free
Get started with Embedding Atlas at no cost with Free version available.
- Free version available
Open Source (MIT)
Open Source (MIT) plan with Full source code on GitHub and Python package (CLI, Jupyter, Streamlit).
- Full source code on GitHub
- Python package (CLI, Jupyter, Streamlit)
- npm UI component library
- Local, private in-browser usage
Capabilities
Key Features
- Interactive embedding visualization in the browser
- Cross-filtering and metadata search
- Local computation: data does not leave your machine
- Python CLI for quick exploration of large datasets
- Jupyter widget and Streamlit component
- WebAssembly UMAP for dimensionality reduction
- Approximate nearest-neighbor search (HNSW, NN-Descent)
- Density-based clustering and automatic labeling
- Load datasets from local files or Hugging Face
- Export standalone web app from CLI