Image Embedding

Generate SigLIP or DINOv2 vectors for image similarity, nearest-neighbor retrieval, and duplicate detection.

Or upload

Upload image

PNG, JPG, WEBP, AVIF (max 50MB)

Embedding Output

Embedding vector payload

Store embeddings in your vector database and use cosine similarity for search and recommendation.

No embedding yet

Submit an image to generate its vector representation.

Use the API directly

Every run maps to POST /api/image/embedding/v1. Generate vectors from URLs or uploads using the same payload in your backend.

Useful when

  • Index images for semantic search and retrieval.
  • Cluster visual assets by similarity.
  • Power recommendation systems using vector distance.