Jina AI Embedding Provider

The Jina AI embedding provider implements the embedding.Embedder interface using the Jina Embeddings API. It supports Jina’s multilingual embedding models with native batch processing and token usage reporting.

Choose Jina embeddings when you need language-specific models (German, Chinese, and others) or when the v3 model’s strong retrieval performance on benchmarks matters for your use case. Jina specializes in search and retrieval, so its models are tuned for similarity search quality.

Installation

go get github.com/lookatitude/beluga-ai/rag/embedding/providers/jina

Quick Start

package main

import (
    "context"
    "fmt"
    "log"
    "os"

    "github.com/lookatitude/beluga-ai/config"
    "github.com/lookatitude/beluga-ai/rag/embedding"
    _ "github.com/lookatitude/beluga-ai/rag/embedding/providers/jina"
)

func main() {
    emb, err := embedding.New("jina", config.ProviderConfig{
        APIKey: os.Getenv("JINA_API_KEY"),
    })
    if err != nil {
        log.Fatal(err)
    }

    ctx := context.Background()
    vectors, err := emb.Embed(ctx, []string{
        "Beluga AI provides a unified embedding interface",
        "Jina AI specializes in search and embeddings",
    })
    if err != nil {
        log.Fatal(err)
    }

    fmt.Printf("Embedded %d texts, dimensions: %d\n", len(vectors), emb.Dimensions())
}

Configuration

Parameter	Type	Default	Description
APIKey	string	(required)	Jina AI API key
Model	string	jina-embeddings-v2-base-en	Embedding model name
BaseURL	string	https://api.jina.ai/v1	API base URL
Timeout	time.Duration	0 (no timeout)	Request timeout
Options["dimensions"]	float64	Model-dependent	Override vector dimensions

Supported Models

Model	Default Dimensions
jina-embeddings-v2-base-en	768
jina-embeddings-v2-small-en	512
jina-embeddings-v2-base-de	768
jina-embeddings-v2-base-zh	768
jina-embeddings-v3	1024

Direct Construction

import (
    jinaemb "github.com/lookatitude/beluga-ai/rag/embedding/providers/jina"
)

emb, err := jinaemb.New(config.ProviderConfig{
    APIKey: os.Getenv("JINA_API_KEY"),
    Model:  "jina-embeddings-v3",
})
if err != nil {
    log.Fatal(err)
}

Multilingual Models

Jina offers language-specific models that can be selected via the model parameter:

// German-language embeddings
emb, err := embedding.New("jina", config.ProviderConfig{
    APIKey: os.Getenv("JINA_API_KEY"),
    Model:  "jina-embeddings-v2-base-de",
})

// Chinese-language embeddings
emb, err := embedding.New("jina", config.ProviderConfig{
    APIKey: os.Getenv("JINA_API_KEY"),
    Model:  "jina-embeddings-v2-base-zh",
})

Batch Embedding

The Jina provider sends all texts in a single API request and correctly maps results back by index:

texts := []string{
    "First document",
    "Second document",
    "Third document",
}

vectors, err := emb.Embed(ctx, texts)
if err != nil {
    log.Fatal(err)
}