Embedding Atlas
ActiveDescription
An interactive visualization tool for large embeddings by Apple. Explore, cross-filter, and search embeddings and metadata to understand and debug embedding models, vector retrieval, and RAG system behavior.
Key Features
- Interactive visualization for large embeddings with automatic clustering and labeling
- Kernel density estimation and density contours for exploring data distributions
- Real-time search and nearest neighbor discovery on embedding data
- Smooth performance at scale — up to millions of points via WebGPU
- Linked dashboards with cross-filtering for tabular data analysis
- MCP support — AI agents can query schema, run SQL, and create charts
Use Cases
💡 Debugging and understanding embedding model behavior on large datasets
💡 Visualizing vector retrieval results and similarity clusters
💡 Exploring RAG system embeddings to identify retrieval quality issues
💡 Analyzing tabular data with linked charts and multimodal viewers
💡 Using MCP to let AI agents explore and visualize data interactively
Categories
Quick Start
```bash
# Install via pip
pip install embedding-atlas
embedding-atlas your-dataset
# Or use in Jupyter
python -c "
from embedding_atlas.widget import EmbeddingAtlasWidget
EmbeddingAtlasWidget(df)
"
# Or install via npm
npm install embedding-atlas
```