Dense 384-dim vectors
Each text becomes a fixed 384-dimensional vector, normalized so cosine similarity is a simple dot product.
CatalogLatency Typically 0.2–1 second (batches are faster per text) Async false Rate Limit Per API key Max Input Many texts per request; very long texts are trimmed automatically
cn8/text-embeddings
Send a list of texts and get back one 384-dimensional vector per text, along with
Learn more about Text Embeddings Text NLPsyncInstant previewtextanalysisnlp
Try it
Texts to embedOne text per line — each becomes a vector.
Result
Results show up here
The sample input is ready — just hit Run.
Overview
Send a list of texts and get back one 384-dimensional vector per text, along with the vector dimensions and the model identifier. Use the vectors for semantic search indexing, clustering, duplicate detection, or as features for your own ML. Responses are synchronous, and you can embed many texts in a single request.
Pricing is per request, not per text — embedding 100 texts in one call costs the same $0.002 as embedding one.
Key capabilities
Batch in one request
Embed any number of texts in a single call and pay one per-request fee.
Multilingual
Texts in different languages are mapped into the same vector space.
Self-describing response
Each response includes the vector dimensions and model identifier so you can sanity-check vectors before storing them.
When to use it
Search indexing
Embed documents and store the vectors in a vector database for similarity search.
Clustering
Group documents, users, or products by vector similarity.
Duplicate detection
Compare vectors to find near-duplicate content.
Feature input
Use embeddings as features for classification, recommendation, or other ML pipelines.
Input & output
input
A texts array of strings.
JSON body
output
One vector per text, plus the vector dimensions and model identifier.
JSON
Guides & tips
How it works
- Each text is converted into a fixed 384-dimensional vector, in the same order as your input array.
- Vectors are normalized, so cosine similarity between two vectors is just their dot product.
embeddingsis always a list of vectors —embeddings[i]is the vector fortexts[i]— even when you send a single text.
Tips
- Cost is per request, so batch as many texts per call as your payload allows.
- For similarity comparisons, cosine similarity (the convention) and dot product are equivalent here since vectors are normalized.
- Store the vectors and reuse them — they're stable across calls for the same text.
Specs
Schema
Request body
textsrequiredarray<string>The texts to embed. Always send an array — even one text needs a one-element array.
Response
embeddingsarray<array<number>>A list of vectors. embeddings[i] corresponds to texts[i] in your request. Each vector has 384 floats.
dimensionsintegerThe length of each vector (384).
modelstringIdentifier of the embedding model used to produce the vectors.
Pricing
Billed per request.
| Service | Unit | Price |
|---|---|---|
| Text Embeddings | item | $0.002/request |
Cost is per request, not per text. Batch as many texts as fit in your payload to make the most of each call.