jina-embeddings-v5-omni-nano

Name: jina-embeddings-v5-omni-nano
Author: jinaai

by jinaai

Compact omni-modal embedding model for text, images, video, and audio in one vector space

8Kdl/month

40likes

986Mparams

HuggingFace Run it on your own data, free

Identifiers

Model ID

jinaai/jina-embeddings-v5-omni-nano

Feature URI

mixpeek://image_extractor@v1/jina_embeddings_v5_omni_nano

Overview

Jina Embeddings v5 Omni Nano is the smallest model in the Jina v5 omni family, placing text, images, video frames, and audio into a single shared vector space. At ~239M parameters, it runs efficiently on edge devices and high-throughput pipelines.

The model shares the same text embedding space as jina-v5-text, meaning existing text indexes remain backwards-compatible when adding multimodal content. This makes it the lowest-friction path to cross-modal search.

Architecture

Multimodal transformer encoder with separate input projections for text, image, video, and audio modalities. All modalities project into a shared embedding space. Matryoshka representation learning enables flexible output dimensions.

Mixpeek SDK Integration

// No extractor parameter takes a Hugging Face model id (checked against
// GET /v1/discovery/extractors, which returns 13), so jina-embeddings-v5-omni-nano runs
// on your side and the output is upserted through POST
// /v1/namespaces/{namespace_id}/documents/upsert. On Enterprise the other
// path is to upload the weights instead: POST /v1/namespaces/{id}/models
// accepts the huggingface format and a custom plugin loads them.
const res = await fetch(
  "https://api.mixpeek.com/v1/namespaces/ns_your_namespace/documents/upsert",
  {
    method: "POST",
    headers: {
      Authorization: "Bearer API_KEY",
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      collection_id: "col_your_collection",
      documents: [
        {
          document_id: "asset-00412",
          // The vector name has to match a vector index on the collection.
          vectors: { "image-embedding": yourVector },
          payload: { source_key: "archive/2026/asset-00412" },
        },
      ],
    }),
  },
);

// Managed alternative, if this exact model is not the requirement:
// image_extractor@v1 runs google/siglip-base-patch16-224
// (768-d) over a bucket, with no inference of your own.