Skip to content
Nappai Documentation

Azure OpenAI Embeddings

This component helps Nappai understand the meaning of text by converting it into numbers. Think of it as translating words into a language that computers can easily understand and work with. This is useful for many tasks within Nappai, such as finding similar documents or categorizing information.

Relationship with Azure OpenAI

This component uses Azure OpenAI’s powerful language models to create these numerical representations (embeddings) of your text. It directly connects to your Azure OpenAI service to perform this conversion.

Inputs

  • Azure Endpoint: The web address of your Azure OpenAI service. This is like the address of your online account. (Required)
  • Deployment Name: The name of your specific deployment within Azure OpenAI. This identifies which model to use. (Required)
  • Credential: Your access key for your Azure OpenAI service. This is like your password, allowing Nappai to access your account. (Required)
  • API Version: (Optional) The version of the Azure OpenAI API to use. Usually, you can leave this at the default setting.
  • Dimensions: (Optional, Advanced) The size of the numerical representation. Leave this at the default unless you have specific requirements.

Outputs

The component produces Embeddings, which are numerical representations of your text. These embeddings are then used by other Nappai components to perform tasks like:

  • Finding similar documents: Documents with similar meanings will have similar embeddings.
  • Categorizing information: Embeddings can help group similar pieces of text together.
  • Improving search results: Searching based on embeddings provides more relevant results than traditional keyword searches.

Usage Example

Imagine you have a collection of customer reviews. You can use this component to create embeddings for each review. Then, you can use another Nappai component (like a vector database) to find reviews that are similar in sentiment or topic.

Templates

This component is used in many Nappai workflows, particularly those involving semantic search and text analysis. You can find it in templates related to:

  • Document similarity analysis
  • Customer feedback analysis
  • Knowledge base search

(Specific template names will be shown in the Nappai interface.)

  • Semantic Text Splitter: This component breaks down large texts into smaller, more manageable chunks before creating embeddings.
  • Couchbase, Upstash, Chroma DB, Weaviate, Vectara, Redis, PGVector, FAISS, Astra DB, Qdrant, Pinecone, MongoDB Atlas, Milvus, Supabase, Cassandra: These are vector databases that store and search embeddings efficiently.
  • Text Embedder: Another embedding component, potentially using a different model.

Tips and Best Practices

  • Ensure your Azure OpenAI service is properly configured and accessible.
  • Use the default settings for API Version and Dimensions unless you have a specific reason to change them.
  • Consider using a semantic text splitter for very long texts to improve accuracy and efficiency.

Security Considerations

Protect your Azure OpenAI credentials carefully. Do not share them with unauthorized individuals. Nappai employs security measures to protect your data, but proper credential management is crucial.