Jina Embedder

The JinaEmbedder class is used to embed text data into vectors using the Jina AI API. You can get started with Jina AI here.

Get your API key.

Usage

jina_embedder.py

from agno.knowledge.knowledge import Knowledge
from agno.vectordb.pgvector import PgVector
from agno.knowledge.embedder.jina import JinaEmbedder

# Add embedding to database
embeddings = JinaEmbedder(id="jina-embeddings-v3").get_embedding("The quick brown fox jumps over the lazy dog.")
# Print the embeddings and their dimensions
print(f"Embeddings: {embeddings[:5]}")
print(f"Dimensions: {len(embeddings)}")

# Use an embedder in a knowledge base
knowledge = Knowledge(
    vector_db=PgVector(
        db_url="postgresql+psycopg://ai:ai@localhost:5532/ai",
        table_name="jina_embeddings",
        embedder=JinaEmbedder(id="jina-embeddings-v3"),
    ),
    max_results=2,
)

Advanced Usage

# Configure embedder with custom settings
embedder = JinaEmbedder(
    id="jina-embeddings-v3",
    dimensions=1024,
    embedding_type="float",
    late_chunking=True,
    batch_size=50,
    timeout=30.0
)

# Use async methods for better performance
import asyncio

async def embed_texts():
    embedder = JinaEmbedder()
    texts = ["First text", "Second text", "Third text"]
    
    # Get embeddings in batches
    embeddings, usage = await embedder.async_get_embeddings_batch_and_usage(texts)
    print(f"Generated {len(embeddings)} embeddings")
    print(f"Usage info: {usage[0]}")

# Run async example
asyncio.run(embed_texts())

Params

Parameter	Type	Default	Description
id	str	"jina-embeddings-v3"	The model ID to use for generating embeddings.
dimensions	int	1024	The number of dimensions for the embedding vectors.
embedding_type	Literal["float", "base64", "int8"]	"float"	The format type of the returned embeddings.
late_chunking	bool	False	Whether to enable late chunking optimization.
user	Optional[str]	None	User identifier for tracking purposes. Optional.
api_key	Optional[str]	JINA_API_KEY env var	The Jina AI API key. Can be set via environment variable.
base_url	str	"https://api.jina.ai/v1/embeddings"	The base URL for the Jina API.
headers	Optional[Dict[str, str]]	None	Additional headers to include in API requests. Optional.
request_params	Optional[Dict[str, Any]]	None	Additional parameters to include in the API request. Optional.
timeout	Optional[float]	None	Timeout in seconds for API requests. Optional.
enable_batch	bool	False	Enable batch processing to reduce API calls and avoid rate limits
batch_size	int	100	Number of texts to process in each API call for batch operations.

Features

• Async Support: Full async/await support for better performance in concurrent applications
• Batch Processing: Efficient batch processing of multiple texts with configurable batch size
• Late Chunking: Support for Jina’s late chunking optimization technique
• Flexible Output: Multiple embedding formats (float, base64, int8)
• Usage Tracking: Get detailed usage information for API calls
• Error Handling: Robust error handling with fallback mechanisms

Developer Resources

• View Cookbook
• Jina AI Documentation