Skip to content
Nappai Documentation

Ollama Embeddings

This component helps Nappai understand the meaning of text by converting it into numbers. These numbers, called “embeddings,” allow Nappai to perform tasks like searching, comparing, and categorizing information more effectively. Think of it as giving Nappai a way to “understand” the meaning behind your text data.

Relationship with Ollama

This component connects to the Ollama AI service to generate these text embeddings. Ollama provides powerful language models that understand the nuances of text. This component acts as a bridge, allowing Nappai to leverage Ollama’s capabilities.

Inputs

  • Ollama Model: Specifies which Ollama model to use for creating the embeddings. The default is a good starting point, but you might explore other models for different results.
  • Ollama Base URL: This is the web address where the Ollama service is running. Usually, you won’t need to change this. The default is typically correct.
  • Model Temperature (Advanced): Controls how creative the model is. A lower value (closer to 0) makes the output more consistent and predictable. A higher value introduces more randomness. Leave this at the default unless you need to fine-tune the results.

Outputs

The component produces Embeddings. These are numerical representations of your text. They are then used by other Nappai components, such as vector databases, to perform tasks like finding similar text or classifying documents.

Usage Example

Imagine you want to find documents similar to a specific customer query. You would feed the query text into the Ollama Embeddings component. The component would generate embeddings, which are then passed to a vector database component (like Pinecone or Weaviate). The database then quickly finds the most similar documents based on their embeddings.

Templates

This component is used in many Nappai workflows, but its specific configuration depends on the overall automation task. You’ll find it used in templates related to semantic search and text analysis.

  • Semantic Text Splitter: This component breaks down large texts into smaller, meaningful chunks before embedding, improving accuracy.
  • Couchbase, Upstash, Chroma DB, Weaviate, Vectara, Redis, PGVector, FAISS, Astra DB, Qdrant, Pinecone, MongoDB Atlas, Milvus, Supabase, Cassandra, Text Embedder: These are all vector databases that store and search embeddings generated by this component. They allow for efficient searching and retrieval of similar text.

Tips and Best Practices

  • Start with the default settings. Adjust the “Model Temperature” only if you need more or less variability in the embeddings.
  • Ensure your Ollama service is running and accessible before using this component.
  • If you encounter errors, double-check the “Ollama Base URL” to make sure it’s correct.

Security Considerations

Ensure your Ollama API is secured according to Ollama’s security best practices. The URL you provide should only point to a trusted and secure Ollama instance.