Embeddings

Preview

Turn text into vectors for semantic search, clustering and retrieval, priced per input token only.

Last updated 2026-07-24

Preview: This endpoint ships with the API launch. The shape below is the interface we are building to. It is documented now so you can plan an integration, not so you can call it today.

An embedding is a numeric representation of meaning. Texts with similar meaning produce vectors that sit close together, which is what makes semantic search, deduplication, clustering and retrieval-augmented generation possible.

Embedding models differ from chat models in two practical ways: they are priced on input only, since there is no generated output, and their dimension count determines both storage cost and how your vector database is configured. Changing embedding model later means re-embedding your whole corpus. This is a decision worth making carefully.

Creating embeddings

curl https://api.webparam.com/v1/embeddings \
  -H "Authorization: Bearer $WEBPARAM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "qwen/qwen3-embedding",
    "input": ["Context windows measure how much a model can read at once."]
  }'

Illustrative: API not yet live

Note: Embedding models appear in the catalogue under the Embeddings category, where pricing is shown as a single input rate rather than the usual input/output pair.

While the API is in build, you can explore the same models two ways: ask Echo how a request and response fit together, or browse the catalogue to compare pricing, context and capabilities.