Add Jina v5 embeddings provider

[hanxiao]

Summary

Add Jina AI as a supported embedding provider for Graphiti knowledge graphs.

jina-embeddings-v5-text-nano (239M params) is the top-ranked embedding model under 500M parameters on MTEB:

• MTEB English v2: 71.0 avg
• MMTEB multilingual: 65.5 avg
• Half the size of comparable models with superior performance across retrieval, STS, and reranking tasks

MMTEB scores vs model size. jina-v5-text models (red) outperform models 2-16x their size. (source)

MTEB English v2 scores. v5-text-nano (239M) achieves 71.0, matching models with 2x+ parameters. (source)

Model	Params	Dim	Max Tokens	MTEB-EN	MMTEB
jina-embeddings-v5-text-nano (default)	239M	768	8192	71.0	65.5
jina-embeddings-v5-text-small	677M	1024	32768	71.7	67.0

Paper: arXiv:2602.15547 | Blog | HuggingFace

Changes

• New JinaEmbedder and JinaEmbedderConfig in graphiti_core/embedder/jina.py
• Uses Jina's OpenAI-compatible API (https://api.jina.ai/v1/embeddings)
• Supports task-specific embeddings via task parameter (retrieval.passage, retrieval.query, etc.)
• API key via api_key config or JINA_API_KEY environment variable
• Default model: jina-embeddings-v5-text-nano (768d)
• Also supports: jina-embeddings-v5-text-small (1024d)
• Comprehensive unit tests with mocked API calls

Usage

from graphiti_core.embedder import JinaEmbedder, JinaEmbedderConfig

# Default configuration
config = JinaEmbedderConfig(
    api_key="your-jina-api-key",
    embedding_model="jina-embeddings-v5-text-nano",
    embedding_dim=768,
    task="retrieval.passage"
)
embedder = JinaEmbedder(config)

# For queries
query_config = JinaEmbedderConfig(
    api_key="your-jina-api-key",
    task="retrieval.query"
)
query_embedder = JinaEmbedder(query_config)

[danielchalef]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[hanxiao]

I have read the CLA Document and I hereby sign the CLA