Embeddings API


POST https://api.fenno.ai/v1/embeddings
Authorization: Bearer <FENNO_API_KEY>
Content-Type: application/json

Request fields

Field	Type	Meaning
`model`	string, required	An available embedding model, not a chat model
`input`	string or array, required	Text or a supported batch of inputs
`dimensions`	integer, optional	Only for models accepting dimension reduction
`encoding_format`	string, optional	A supported format such as `float`; do not assume every model supports alternatives

Models impose their own single-input, total-token and batch limits. Split oversized content in the application; do not silently truncate text needed for retrieval.

Python example

Install openai; set FENNO_API_KEY and FENNO_EMBEDDING_MODEL:


import os
from openai import OpenAI
 
client = OpenAI(
    base_url="https://api.fenno.ai/v1",
    api_key=os.environ["FENNO_API_KEY"],
    timeout=60.0,
    max_retries=0,
)
 
response = client.embeddings.create(
    model=os.environ["FENNO_EMBEDDING_MODEL"],
    input=["Database backup guide", "Database restore guide"],
)
for item in response.data:
    print(item.index, len(item.embedding))

Response

data[] contains vectors with index, object and embedding; model identifies the model and usage reports token usage when provided. Use index to map vectors back to input records.

For float encoding, embedding is an array of numbers. If another encoding is selected, decode it according to that encoding rather than passing it directly into a vector index.

Do not mix vectors from different models or dimensions. Store the embedding configuration with the index and use it consistently for queries. See semantic retrieval for the full workflow.