Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/dais-polymtl/flock/llms.txt

Use this file to discover all available pages before exploring further.

Flock supports Azure OpenAI as a dedicated provider. You connect Flock to your Azure deployment by storing your API key, resource name, and API version as a DuckDB secret, then registering a model that maps to a specific deployment. All standard Flock functions work identically once the provider is configured.

​
Prerequisites

You need the following from your Azure OpenAI resource:
  • API key — found under Keys and Endpoint in the Azure portal
  • Resource name — the subdomain of your endpoint (for example, my-resource if your endpoint is https://my-resource.openai.azure.com)
  • API version — the Azure OpenAI REST API version string (for example, 2024-02-01)
  • Deployment name — the name you gave to the model deployment in Azure AI Studio
For the full Azure OpenAI reference, see the Microsoft Azure documentation. Make sure Flock is installed and loaded before continuing — see the Quickstart if you haven’t done that yet.

​
Configure your secret

Store your Azure credentials using CREATE SECRET with the azure_llm type: 
CREATE SECRET (
    TYPE azure_llm,
    API_KEY 'your-api-key',
    RESOURCE_NAME 'your-resource-name',
    API_VERSION 'your-api-version'
);
All three fields are required. Flock reconstructs the endpoint URL as https://{RESOURCE_NAME}.openai.azure.com internally.
RESOURCE_NAME is just the subdomain, not the full URL. If your Azure endpoint is https://my-resource.openai.azure.com, set RESOURCE_NAME to my-resource.

​
Create a model

Register a named model pointing to your Azure deployment. The second argument is your deployment name as configured in Azure AI Studio: 
CREATE MODEL(
    'QuackingModel',
    'gpt-4o',
    'azure',
    {"tuple_format": "json", "batch_size": 32, "model_parameters": {"temperature": 0.7}}
);
The four arguments are:
ArgumentDescription
'QuackingModel'Unique name you reference in queries
'gpt-4o'Azure deployment name
'azure'Provider name
{...}Config: batch size, tuple format, and model parameters

​
Run a query

With your secret and model in place, call llm_complete: 
SELECT llm_complete(
    {'model_name': 'QuackingModel'},
    {'prompt': 'Write a short poem about a database.'}
);
To use column data as context: 
SELECT llm_complete(
    {'model_name': 'QuackingModel'},
    {
        'prompt': 'Summarize this document in one paragraph: {doc}',
        'context_columns': [{'data': document_text, 'name': 'doc'}]
    }
) AS summary
FROM documents;

​
Azure-specific configuration details

Flock builds the Azure endpoint from RESOURCE_NAME automatically:
https://{RESOURCE_NAME}.openai.azure.com
You do not need to supply the full URL — only the resource name subdomain.
Azure OpenAI requires an explicit API version with every request. Pass it as API_VERSION when creating your secret. Common values include 2024-02-01 and 2024-05-01-preview. Check the Azure OpenAI API reference for the current stable version.
In Azure, you deploy a base model (such as gpt-4o) under a custom deployment name you choose. The second argument to CREATE MODEL should be your deployment name, not necessarily the underlying model ID — they may differ depending on how you named your deployment.

​
Next steps

Structured output

Enforce JSON schemas on llm_complete responses

Scalar functions

Full reference for llm_complete, llm_filter, llm_embedding, and more

Build docs developers (and LLMs) love

Get started for freeTalk to us