Nice work! Desktop apps for DB exploration always feel more responsive than web alternatives, especially when you're iterating on embeddings and need quick feedback.
A few questions: does it support visualizing vector clusters (like t-SNE or UMAP projections)? And is there a way to test similarity queries interactively before committing them to code?
Vector DBs have this unique challenge where the data is high-dimensional and hard to intuit. Tools that help developers 'see' what's happening in the embedding space are incredibly valuable.
Really appreciate the commitment to client-side processing. The ONNX + WASM fallback approach is smart for broader compatibility.
Curious about PDF tools - any plans to add PDF compression or manipulation? That's one area where most "free" tools are especially sketchy about data handling. Client-side PDF processing seems like a natural fit for this philosophy.
Also wondering: have you considered adding a visual indicator (like a network activity monitor) to reassure users that nothing is being uploaded? Could reinforce the privacy promise.
A few questions: does it support visualizing vector clusters (like t-SNE or UMAP projections)? And is there a way to test similarity queries interactively before committing them to code?
Vector DBs have this unique challenge where the data is high-dimensional and hard to intuit. Tools that help developers 'see' what's happening in the embedding space are incredibly valuable.
reply