Jina AI Embeddings API

Jina AI Embeddings API is a token metered service for generating fixed length vector representations from text and images through the endpoint https://api.jina.ai/v1/embeddings. You send inputs in a JSON request and receive embeddings that can be stored in a vector database for search or retrieval. It exposes practical controls for downstream retrieval. The API supports optional L2 normalization and lets you choose an embedding output type such as float, binary, or base64 for faster retrieval or transmission, depending on your stack. Access is managed by an API key that also applies across Jina Search Foundation products, with token usage shared. The site describes rate limits tracked by requests per minute and tokens per minute, enforced per IP or per key when provided, with higher limits available for paid tiers. Jina highlights integrations with common vector stores and RAG frameworks and lists connectors for systems such as MongoDB, Qdrant, Pinecone, Chroma, Weaviate, Milvus and DataStax. For deployment control, it also mentions options to run embeddings on AWS SageMaker, Microsoft Azure, Google Cloud marketplaces, plus custom Kubernetes deployments for VPC or on premises. Pricing is expressed in tokens with a free trial per new API key, and additional token packages can be purchased via top ups and optional auto recharge when balances fall below a threshold. A licensing self check is provided for commercial use scenarios.