Inference
Embed
Generates embeddings for text, image, audio, or video data to create machine-readable representations.
POST
/
embed
Request
model_id
string
requiredThe model_id to use for embedding,
input
string
requiredThe input source for the file to be embedded, e.g., a URL or direct content.
Max of 5mb per request can be sent (via url or base64), consider using the File Tools API to pre-process larger inputs
modality
string
requiredThe type of data to be embedded (e.g., text, image, audio, video).
input_type
string
requiredThe type of input provided (e.g., “url” for remote URLs, “text” for text or “base64” for base64-encoded content).