Link Search Menu Expand Document Documentation Menu
Documentation

Supported connectors

OpenSearch provides two types of connector blueprints:

  • Standard blueprints (recommended for OpenSearch 2.14+):

    • Designed for connectors that pass input directly to the model and return its raw output without requiring additional transformations.
    • Use a registered model without any pre-processing or post-processing functions.
    • Compatible with machine learning (ML) inference processors: ingest processors, search request processors, and search response processors.
    • Ideal for new implementations in OpenSearch 2.14 and later.

  • Legacy blueprints:

    • Suitable for specific input and output formatting requirements.
    • Include pre- and post-processing functions.
    • Recommended for existing implementations.
    • Compatible with neural search queries.

The following table provides a comprehensive list of connector blueprints available in OpenSearch. Each blueprint link offers an example request for creating the connector, including all necessary parameters, and a sample Predict API request.

Platform Model Legacy blueprint Standard blueprint
Aleph Alpha Luminous-Base Embedding Model Blueprint N/A
Amazon Bedrock AI21 Labs Jurassic-2 Mid Blueprint Blueprint
Amazon Bedrock Anthropic Claude v2 Blueprint Blueprint
Amazon Bedrock Anthropic Claude v3 Blueprint Blueprint
Amazon Bedrock Anthropic Claude v3.7 Blueprint Blueprint
Amazon Bedrock Anthropic Claude v4 Blueprint Blueprint
Amazon Bedrock Cohere Embed Model v3 - English Blueprint Blueprint
Amazon Bedrock Cohere Embed Model v3 - Multilingual Blueprint Blueprint
Amazon Bedrock Titan Text Embeddings Blueprint Blueprint
Amazon Bedrock Titan Multimodal Embeddings Blueprint Blueprint
Amazon Bedrock Nova Multimodal Embeddings Blueprint Blueprint
Amazon Bedrock OpenAI GPT-OSS Blueprint Blueprint
Amazon Bedrock Converse Anthropic Claude 3 Sonnet Blueprint Blueprint
Amazon Comprehend Metadata Embedding Model Blueprint Blueprint
Amazon Textract Amazon Textract Model Blueprint Blueprint
Amazon SageMaker Text embedding models for batch inference Blueprint Blueprint
Amazon SageMaker Text embedding models Blueprint Blueprint
Amazon SageMaker Semantic highlighter models N/A Blueprint
Azure OpenAI Chat models (for example, gpt-4) Blueprint Blueprint
Azure OpenAI Text embedding models (for example, text-embedding-ada-002) Blueprint Blueprint
Cohere Chat models Blueprint Blueprint
Cohere Multimodal embedding models (for example, embed-english-v3.0) Blueprint Blueprint
Cohere Text embedding models Blueprint Blueprint
DeepSeek Chat model Blueprint Blueprint
Google Cloud Platform VertexAI Embedding Model Blueprint Blueprint
Ollama Chat models Blueprint Blueprint
OpenAI Chat models (for example, gpt-3.5-turbo) Blueprint Blueprint
OpenAI Completion models (for example, text-davinci-003) Blueprint Blueprint
OpenAI Text embedding models (for example, text-embedding-ada-002) Blueprint Blueprint
OpenAI Text embedding models for batch inference Blueprint Blueprint
Yandex Cloud Text embedding models Blueprint Blueprint

LLM judgment blueprints for Search Relevance Workbench

Search Relevance Workbench uses a chat model to generate relevance ratings using an LLM-as-a-Judge. For an overview of these blueprints, see the LLM judgment blueprints README. The following table lists the available blueprints for connecting Search Relevance Workbench to a chat model.

Provider Blueprint
OpenAI GPT-5, GPT-4.1, GPT-4o
Azure OpenAI GPT-5, GPT-4.1, GPT-4o
DeepSeek DeepSeek-V3, DeepSeek-R1
Ollama and other local OpenAI-compatible servers Qwen 3, Llama 3.3, Mistral, Phi-4, Gemma 3
Google Gemini Gemini 2.5 Pro, Gemini 2.5 Flash
Amazon Bedrock Anthropic Claude (native invoke API)
Amazon Bedrock Converse Anthropic Claude, Amazon Nova, Meta Llama, Mistral, AI21, Cohere

For more information about using these blueprints with SRW, see Using different LLM providers.

350 characters left

Have a question? .

Want to contribute? or .