OpenAI. The selected embedding model provider for this configuration.

Required. API key for the OpenAI API. If your product allows runtime injection, follow your admin guide for leaving this blank. Stored keys must be protected like any other secret.

Required. Embedding model id (for example text-embedding-3-small, text-embedding-3-large, or text-embedding-ada-002). Use the exact id your account and API version support.

Optional. Overrides the default OpenAI API base URL. Leave blank to use the standard endpoint unless you use a proxy or compatible gateway.

Optional. Desired embedding vector length when the model supports multiple output dimensions (for example some text-embedding-3-* models). Leave blank to use the model default. Must be valid for the chosen model.

Optional. OpenAI organization id for usage and access scoping. Leave blank if you do not use organizations.

Optional. OpenAI project id when your workflow uses project-scoped keys or reporting. Leave blank if not used.

Optional. End-user or caller identifier OpenAI accepts for abuse monitoring and analytics (stable id per end user where policy allows). Leave blank if not required.

Optional. Maximum number of texts to send per embedding API call. Larger batches can improve throughput but hit payload and rate limits; leave blank to use the product or model default.

Optional. Number of times to retry a failed request after errors such as rate limits or transient network failures. Default is often 3 when not set.

Optional. Maximum wait time for a single request, in milliseconds. Default is often 60000 (60 seconds) when not set. Increase for large batches or slow networks.

When enabled, request details are written to logs. May include text sent for embedding—use only while debugging and disable afterward.

When enabled, response details are written to logs. May include vectors and metadata—use only while debugging and disable afterward.

Azure OpenAI. The selected embedding model provider for this configuration.

Required. Key for your Azure OpenAI resource (from the Azure portal or your secret store). Stored keys must be protected.

Required. Name of the embedding deployment in Azure OpenAI (Azure AI Foundry / Studio). Must match a deployment that uses an embedding-capable model.

Required. Endpoint URL for the Azure OpenAI resource, typically https://<resource-name>.openai.azure.com/ (or the path your gateway uses).

Optional. Output embedding size when the deployed model supports configurable dimensions. Leave blank to use the deployment default.

Optional. Retries for failed requests. Default is often 3 when not set.

Optional. Request timeout in milliseconds. Default is often 60000 when not set.

When enabled, logs request details; embedding inputs may be sensitive—use short-term for troubleshooting only.

When enabled, logs response details; disable when finished debugging.