--- title: "Jina AI Multimodal Embeddings with Weaviate" description: "Weaviate's integration with Jina AI's APIs allows you to access their models' capabilities directly from Weaviate." url: https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-multimodal --- # Jina AI Multimodal Embeddings with Weaviate Weaviate's integration with Jina AI's APIs allows you to access their models' capabilities directly from Weaviate. [Configure a Weaviate vector index](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-multimodal#configure-the-vectorizer) to use a Jina AI embedding model, and Weaviate will generate embeddings for various operations using the specified model and your Jina AI API key. This feature is called the *vectorizer*. At [import time](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-multimodal#data-import), Weaviate generates multimodal object embeddings and saves them into the index. For [vector](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-multimodal#vector-near-text-search) and [hybrid](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-multimodal#hybrid-search) search operations, Weaviate converts queries of one or more modalities into embeddings. [Multimodal search operations](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-multimodal#vector-near-media-search) are also supported. ![Embedding integration illustration](https://docs.weaviate.io/assets/images/integration_jinaai_embedding-2843eca6d973cbf4769c8464427f48be.png) ## Requirements ### Weaviate configuration Your Weaviate instance must be configured with the Jina AI multimodal vectorizer integration (`multi2vec-jinaai`) module. **For Weaviate Cloud (WCD) users** This integration is enabled by default on Weaviate Cloud (WCD) instances. **For self-hosted users** - Check the [cluster metadata](https://docs.weaviate.io/deploy/configuration/status#cluster-metadata) to verify if the module is enabled. - Follow the [how-to configure modules](https://docs.weaviate.io/weaviate/configuration/modules) guide to enable the module in Weaviate. ### API credentials You must provide a valid Jina AI API key to Weaviate for this integration. Go to [Jina AI](https://jina.ai/embeddings/) to sign up and obtain an API key. Provide the API key to Weaviate using one of the following methods: - Set the `JINAAI_APIKEY` environment variable that is available to Weaviate. - Provide the API key at runtime, as shown in the examples below. **Python** ```py import weaviate from weaviate.classes.init import Auth import os # Recommended: save sensitive data as environment variables jinaai_key = os.getenv("JINAAI_API_KEY") headers = { "X-JinaAI-Api-Key": jinaai_key, } client = weaviate.connect_to_weaviate_cloud( cluster_url=weaviate_url, # `weaviate_url`: your Weaviate URL auth_credentials=Auth.api_key(weaviate_key), # `weaviate_key`: your Weaviate API key headers=headers ) # Work with Weaviate client.close() ``` **JavaScript/TypeScript** ```ts import weaviate from '@weaviate/node' const jinaaiApiKey = process.env.JINAAI_API_KEY || ''; // Replace with your inference API key const client = await weaviate.connectToWeaviateCloud( 'WEAVIATE_INSTANCE_URL', // Replace with your instance URL { authCredentials: new weaviate.ApiKey('WEAVIATE_INSTANCE_API_KEY'), headers: { 'X-JinaAI-Api-Key': jinaaiApiKey, } } ) // Work with Weaviate client.close() ``` ## Configure the vectorizer [Configure a Weaviate index](https://docs.weaviate.io/weaviate/manage-collections/vector-config#specify-a-vectorizer) as follows to use a Jina AI embedding model: **Python** ```py from weaviate.classes.config import Configure, DataType, Multi2VecField, Property client.collections.create( "DemoCollection", properties=[ Property(name="title", data_type=DataType.TEXT), Property(name="poster", data_type=DataType.BLOB), ], vector_config=[ Configure.Vectors.multi2vec_jinaai( name="title_vector", # Define the fields to be used for the vectorization - using image_fields, text_fields image_fields=[ Multi2VecField(name="poster", weight=0.9) ], text_fields=[ Multi2VecField(name="title", weight=0.1) ], ) ], # Additional parameters not shown ) ``` **JavaScript/TypeScript** ```ts await client.collections.create({ name: 'DemoCollection', properties: [ { name: 'title', dataType: 'text' as const, }, { name: 'poster', dataType: 'blob' as const, }, ], vectorizers: [ weaviate.configure.vectors.multi2VecJinaAI({ name: 'title_vector', imageFields: [{ name: "poster", weight: 0.9 }], textFields: [{ name: "title", weight: 0.1 }] }, ), ], // Additional parameters not shown }); ``` ### Select a model You can specify one of the [available models](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-multimodal#available-models) for the vectorizer to use, as shown in the following configuration example. **Python** ```py from weaviate.classes.config import Configure, DataType, Multi2VecField, Property client.collections.create( "DemoCollection", properties=[ Property(name="title", data_type=DataType.TEXT), Property(name="poster", data_type=DataType.BLOB), ], vector_config=[ Configure.Vectors.multi2vec_jinaai( name="title_vector", # Define the fields to be used for the vectorization - using image_fields, text_fields image_fields=[ Multi2VecField(name="poster", weight=0.9) ], text_fields=[ Multi2VecField(name="title", weight=0.1) ], model="jina-clip-v2", ) ], ) ``` **JavaScript/TypeScript** ```ts await client.collections.create({ name: 'DemoCollection', properties: [ { name: 'title', dataType: 'text' as const, }, { name: 'poster', dataType: 'blob' as const, }, ], vectorizers: [ weaviate.configure.vectors.multi2VecJinaAI({ name: 'title_vector', imageFields: [{ name: "poster", weight: 0.9 }], textFields: [{ name: "title", weight: 0.1 }], model: "jina-clip-v2" }, ), ], // Additional parameters not shown }); ``` The [default model](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-multimodal#available-models) is used if you do not specify one. **Vectorization behavior** Weaviate follows the collection configuration and a set of predetermined rules to vectorize objects. Unless specified otherwise in the collection definition, the default behavior is to: - Only vectorize properties that use the `text` or `text[]` data type (unless [skipped](https://docs.weaviate.io/weaviate/manage-collections/vector-config#property-level-settings)) - Sort properties in alphabetical (a-z) order before concatenating values - If `vectorizePropertyName` is `true` (`false` by default) prepend the property name to each property value - Join the (prepended) property values with spaces - Prepend the collection name if `vectorizeClassName` is `true` (the default differs by integration — check this page's parameter table) Since `v1.27`, the string is sent to the model as is. Earlier versions lowercase it first, and you can restore that behavior with the [`LOWERCASE_VECTORIZATION_INPUT`](https://docs.weaviate.io/deploy/configuration/env-vars#module-specific) environment variable. ### Vectorizer parameters The following examples show how to configure Jina AI-specific options. - `model`: The model name. - `dimensions`: The number of dimensions for the model. - Note that [not all models](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-multimodal#available-models) support this parameter. **Python** ```py from weaviate.classes.config import Configure, DataType, Multi2VecField, Property client.collections.create( "DemoCollection", properties=[ Property(name="title", data_type=DataType.TEXT), Property(name="poster", data_type=DataType.BLOB), ], vector_config=[ Configure.Vectors.multi2vec_jinaai( name="title_vector", # Define the fields to be used for the vectorization - using image_fields, text_fields image_fields=[ Multi2VecField(name="poster", weight=0.9) ], text_fields=[ Multi2VecField(name="title", weight=0.1) ], # Further options # model="jina-clip-v2", # dimensions=512, # Only applicable for some models (e.g. `jina-clip-v2`) ) ], # Additional parameters not shown ) ``` **JavaScript/TypeScript** ```ts await client.collections.create({ name: 'DemoCollection', properties: [ { name: 'title', dataType: 'text' as const, }, { name: 'poster', dataType: 'blob' as const, }, ], vectorizers: [ weaviate.configure.vectors.multi2VecJinaAI({ name: 'title_vector', imageFields: [{ name: "poster", weight: 0.9 }], textFields: [{ name: "title", weight: 0.1 }], // Further options // model:"jina-clip-v2", }, ), ], // Additional parameters not shown }); ``` ## Data import After configuring the vectorizer, [import data](https://docs.weaviate.io/weaviate/manage-objects/import) into Weaviate. Weaviate generates embeddings for text objects using the specified model. **Python** ```py collection = client.collections.use("DemoCollection") with collection.batch.fixed_size(batch_size=200) as batch: for src_obj in source_objects: poster_b64 = url_to_base64(src_obj["poster_path"]) weaviate_obj = { "title": src_obj["title"], "poster": poster_b64 # Add the image in base64 encoding } # The model provider integration will automatically vectorize the object batch.add_object( properties=weaviate_obj, # vector=vector # Optionally provide a pre-obtained vector ) ``` **JavaScript/TypeScript** ```ts const collectionName = 'DemoCollection' const myCollection = client.collections.use(collectionName) let multiModalObjects = new Array(); for (let mmSrcObject of mmSrcObjects) { multiModalObjects.push({ title: mmSrcObject.title, poster: mmSrcObject.poster, // Add the image in base64 encoding }); } // The model provider integration will automatically vectorize the object const mmResponse = await myCollection.data.insertMany(multiModalObjects); console.log(mmResponse); ``` > [!TIP] > > **Re-use existing vectors** > > If you already have a compatible model vector available, you can provide it directly to Weaviate. This can be useful if you have already generated embeddings using the same model and want to use them in Weaviate, such as when migrating data from another system. ## Searches Once the vectorizer is configured, Weaviate will perform vector and hybrid search operations using the specified Jina AI model. ![Embedding integration at search illustration](https://docs.weaviate.io/assets/images/integration_jinaai_embedding_search-20bbf72e7009131771abd68c5639efcc.png) ### Vector (near text) search When you perform a [vector search](https://docs.weaviate.io/weaviate/search/similarity#search-with-text), Weaviate converts the text query into an embedding using the specified model and returns the most similar objects from the database. The query below returns the `n` most similar objects from the database, set by `limit`. **Python** ```py collection = client.collections.use("DemoCollection") response = collection.query.near_text( query="A holiday film", # The model provider integration will automatically vectorize the query limit=2 ) for obj in response.objects: print(obj.properties["title"]) ``` **JavaScript/TypeScript** ```ts const collectionName = 'DemoCollection' const myCollection = client.collections.use(collectionName) let result; result = await myCollection.query.nearText( 'A holiday film', // The model provider integration will automatically vectorize the query { limit: 2, } ) console.log(JSON.stringify(result.objects, null, 2)); ``` ### Hybrid search > [!NOTE] > > **What is a hybrid search?** > > A hybrid search performs a vector search and a keyword (BM25) search, before [combining the results](https://docs.weaviate.io/weaviate/search/hybrid) to return the best matching objects from the database. When you perform a [hybrid search](https://docs.weaviate.io/weaviate/search/hybrid), Weaviate converts the text query into an embedding using the specified model and returns the best scoring objects from the database. The query below returns the `n` best scoring objects from the database, set by `limit`. **Python** ```py collection = client.collections.use("DemoCollection") response = collection.query.hybrid( query="A holiday film", # The model provider integration will automatically vectorize the query limit=2 ) for obj in response.objects: print(obj.properties["title"]) ``` **JavaScript/TypeScript** ```ts const collectionName = 'DemoCollection' const myCollection = client.collections.use(collectionName) result = await myCollection.query.hybrid( 'A holiday film', // The model provider integration will automatically vectorize the query { limit: 2, } ) console.log(JSON.stringify(result.objects, null, 2)); ``` ### Vector (near media) search When you perform a media search such as a [near image search](https://docs.weaviate.io/weaviate/search/similarity#search-with-image), Weaviate converts the query into an embedding using the specified model and returns the most similar objects from the database. To perform a near media search such as near image search, convert the media query into a base64 string and pass it to the search query. The query below returns the `n` most similar objects to the input image from the database, set by `limit`. **Python** ```py def url_to_base64(url): import requests import base64 image_response = requests.get(url) content = image_response.content return base64.b64encode(content).decode("utf-8") collection = client.collections.use("DemoCollection") query_b64 = url_to_base64(src_img_path) response = collection.query.near_image( near_image=query_b64, limit=2, return_properties=["title", "release_date", "tmdb_id", "poster"] # To include the poster property in the response (`blob` properties are not returned by default) ) for obj in response.objects: print(obj.properties["title"]) ``` **JavaScript/TypeScript** ```ts const base64String = 'SOME_BASE_64_REPRESENTATION'; result = await myCollection.query.nearImage( base64String, // The model provider integration will automatically vectorize the query { limit: 2, } ) console.log(JSON.stringify(result.objects, null, 2)); ``` ## References ### Available models - `jina-clip-v2` (server default) - This model is a multilingual, multimodal model using [Matryoshka Representation Learning](https://arxiv.org/abs/2205.13147). - It will accept a `dimensions` parameter, which can be any integer between (and including) 64 and 1024. The default value is 1024. - `jina-clip-v1` - This model will always return a 768-dimensional embedding. ## Further resources ### Other integrations - [Jina AI text embedding models + Weaviate](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings). - [Jina AI ColBERT embedding models + Weaviate](https://docs.weaviate.io/weaviate/model-providers/jinaai/embeddings-colbert). - [Jina AI reranker models + Weaviate](https://docs.weaviate.io/weaviate/model-providers/jinaai/reranker). ### Code examples Once the integrations are configured at the collection, the data management and search operations in Weaviate work identically to any other collection. See the following model-agnostic examples: - The [How-to: Manage collections](https://docs.weaviate.io/weaviate/manage-collections) and [How-to: Manage objects](https://docs.weaviate.io/weaviate/manage-objects) guides show how to perform data operations (i.e. create, read, update, delete collections and objects within them). - The [How-to: Query & Search](https://docs.weaviate.io/weaviate/search) guides show how to perform search operations (i.e. vector, keyword, hybrid) as well as retrieval augmented generation. ### External resources - Jina AI [Embeddings API documentation](https://jina.ai/embeddings/)