Using your own LLM key for AI usage

Requires an Enterprise license.

These instructions are for Fabricate Cloud only. Self-hosted instances must provide their own LLM key during setup.

On Fabricate Cloud, before you can use this option, you must contact Tonic.ai to enable it for your account.

After Tonic.ai enables the option for you to use your own LLM key, then until you configure the LLM provider and key, Fabricate does not process any LLM requests.

Enterprise accounts can choose to provide an API key to use for Fabricate AI features. The API key must be from one of the following LLM providers:

• Anthropic

• Amazon Bedrock

• Azure AI Foundry

• Google Gemini Enterprise Agent Platform (formerly Vertex AI)

When you provide your own LLM key, Tonic.ai does not charge your for AI usage, only for basic access. However, any configured limit on AI usage cost still applies.

Overview of the flow with your own LLM key

The following diagram provides an overview of the flow of requests and data when you provide your own LLM key, in this case Amazon Bedrock.

Diagram of the flow of requests and data when you use your own LLM key

Enabling the option to provide your own key

To enable the option to use your own LLM key, you must contact Tonic.ai.

After Tonic.ai configures your account to require a custom LLM key, then before you can use Fabricate, you must provide the LLM key.

Selecting the LLM provider

On the AI page of Account Settings, you set up the LLM provider in LLM Provider Settings.

LLM Provider Settings for an Enterprise account

From the Provider dropdown list, select the LLM provider to use.

You can use either:

• Anthropic

• Amazon Bedrock

• Azure AI Foundry

• Google Vertex AI - This is the option for Google Gemini Enterprise Agent Platform

Providing the LLM provider API key

Anthropic

If you choose Anthropic, then in the Anthropic API Key field, provide your Anthropic key.

Amazon Bedrock

If you choose Amazon Bedrock, then:

1. In the Amazon Bedrock API Key field, provide your Amazon Bedrock key.

2. In the AWS Region field, specify the AWS Region where your account is located.

Azure AI Foundry

If you choose Azure AI Foundry, then:

1. In the Azure AI Foundry API Key field, provide the API key for your Azure AI Foundry endpoint.

2. In the Azure AI Foundry Endpoint URL field, provide the base URL for your Azure AI Foundry Anthropic-compatible endpoint.

Google Gemini Enterprise Agent Platform (formerly Vertex AI)

If you choose Google Gemini Enterprise Agent Platform, then:

1. In the Google Cloud Project ID field, provide the identifier of the Google Cloud project where Vertex is enabled.

2. In the Google Cloud Location field, provide the location where the Vertex AI endpoint is deployed.

3. In the Service Account Key (JSON) field, provide the service account JSON key.

OpenAI

If you choose OpenAI, then:

1. In the API key field, provide your API key for OpenAI.

2. Optionally, in the API URL field, provide your custom base URL for the OpenAI API.

Testing and saving the LLM provider key

To test the provided key, click Test Credentials.

To save the configuration, click Save Changes.