Embed text with Qwen3-Embedding

Pick a model from the catalog. Qwen3-Embedding 0.6B is fast and permissive. Passing a catalog id as model_id with no path makes the node fetch the ONNX graph and tokenizer from HuggingFace onto its models directory, then register the encoder under that same id.

tenzro embed-text catalog

tenzro embed-text load --model qwen3-embedding-0.6b

The runtime returns an L2-normalized vector ready for cosine similarity.

tenzro embed-text run \
  --model qwen3-embedding-0.6b \
  --input "the open network for autonomous AI" \
  --normalize

EmbeddingGemma supports compact dims with on-the-fly re-normalization — useful when you store millions of vectors.

tenzro embed-text run \
  --model embeddinggemma-300m \
  --input "compact vector for retrieval" \
  --requested-dim 256 \
  --normalize

The same shape is available over RPC for integration into your own indexers. inputs is an array, so one call embeds a whole batch.

curl -s https://rpc.tenzro.xyz -H 'content-type: application/json' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tenzro_textEmbed","params":{"model_id":"qwen3-embedding-0.6b","inputs":["hello","world"],"normalize":true}}'