Skip to main content
Embeddings convert text into numerical vectors for similarity search.

OpenAI embeddings

from agno.knowledge import Knowledge
from agno.knowledge.embedder.openai import OpenAIEmbedder

knowledge = Knowledge(
    path="docs/",
    vector_db=vector_db,
    embedder=OpenAIEmbedder(model="text-embedding-3-small")
)

Google Gemini embeddings

from agno.knowledge.embedder.google import GeminiEmbedder

knowledge = Knowledge(
    path="docs/",
    vector_db=vector_db,
    embedder=GeminiEmbedder(model="gemini-embedding-001")
)

Azure OpenAI embeddings

from agno.knowledge.embedder.azure import AzureOpenAIEmbedder

knowledge = Knowledge(
    path="docs/",
    vector_db=vector_db,
    embedder=AzureOpenAIEmbedder(
        model="text-embedding-3-small",
        azure_deployment="my-embedding-deployment"
    )
)

Ollama embeddings (local)

from agno.knowledge.embedder.ollama import OllamaEmbedder

knowledge = Knowledge(
    path="docs/",
    vector_db=vector_db,
    embedder=OllamaEmbedder(model="nomic-embed-text")
)
See cookbook examples for patterns.

Build docs developers (and LLMs) love