Skip to content

Latest commit

 

History

History
107 lines (73 loc) · 4.89 KB

connection-strings-overview.markdown

File metadata and controls

107 lines (73 loc) · 4.89 KB

AI Connection Strings - Overview

{NOTE: }

  • In RavenDB, you can define Embeddings Generation Tasks that generate embeddings from the content of your documents. These embeddings are stored in a dedicated collection within the database and enable vector search on your document content.

  • Each embeddings generation task must define a connection string to an embedding provider.
    This connection string specifies where the embeddings will be generated, allowing RavenDB to integrate with external services such as Azure OpenAI, OpenAI, Hugging Face, Google AI, Ollama, Mistral AI, or RavenDB's embedded model (bge-micro-v2).

  • While each task can have only one connection string, you can define multiple connection strings in your database to support different providers or configurations.
    A single connection string can also be reused across multiple tasks in the database.

  • These connection strings can be created from:

    • The AI Connection Strings view in the Studio, where you can create, edit, and delete connection strings that are not in use.
    • The Client API - examples are available in the dedicated articles for each provider.

{NOTE/}

{PANEL: The AI Connection Strings view}

connection strings view

  1. Go to the AI Hub menu.

  2. Open the AI Connection Strings view.

  3. Click "Add new" to create a new connection string.

  4. View the list of all AI connection strings.

  5. Edit or delete a connection string.
    Only connection strings that are not in use by a task can be deleted.

{PANEL/}

{PANEL: Creating an AI connection string}

create connection string

  1. Name
    Enter a unique name for the connection string.

  2. Identifier
    Enter a unique identifier for the connection string.
    Each AI connection string in the database must have a distinct identifier.

    If not specified, or when clicking the "Regenerate" button,
    RavenDB automatically generates the identifier based on the connection string name. For example: * If the connection string name is: "My connection string to Google AI" * The generated identifier will be: "my-connection-string-to-google-ai"

    Allowed characters: only lowercase letters (a-z), numbers (0-9), and hyphens (-).
    See how this identifier is used in the embeddings cache collection.

  3. Regenerate
    Click "Regenerate" to automatically create an identifier based on the connection string name.

  4. Connector
    Select an AI provider from the dropdown menu.
    This will open a popup where you can configure the connection details.
    Configuration details for each provider are explained in the following articles:

  5. Once you complete all configurations for the selected provider in the popup view,
    save the connection string definition.

{PANEL/}

Related Articles

Vector Search

Embeddings Generation

AI Connection Strings