Skip to content
4ALL API Docs
Main Site News Console

Text Embeddings

POST https://api.4allapi.com/v1/embeddings

Converts text into high-dimensional vectors for semantic search, clustering, recommendations, and RAG knowledge bases.

Request Example

Terminal window
curl https://api.4allapi.com/v1/embeddings \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AIPROXY_KEY" \
  -d '{
    "model": "text-embedding-3-large",
    "input": ["What is 4ALL API", "How to create an API token"]
  }'

input supports a single string or an array of strings (batch); each input in the response corresponds to one embedding vector:

{
  "object": "list",
  "data": [
    { "object": "embedding", "index": 0, "embedding": [0.0123, -0.0456, ...] },
    { "object": "embedding", "index": 1, "embedding": [...] }
  ],
  "model": "text-embedding-3-large",
  "usage": { "prompt_tokens": 14, "total_tokens": 14 }
}

Practical Recommendations

  • Use the same embedding model for the entire knowledge base, and do not mix vector spaces from different models;
  • For long documents, split into chunks first (e.g. 300–800 characters, with overlap) before embedding; this improves retrieval hit rates;
  • When embedding in batches, submit multiple items at once as an array; this is faster and more efficient than one request at a time;
  • Embeddings are billed by input tokens, and the price is far lower than chat models, making them suitable for large-scale indexing.