Create your first LLM key

Overview

The following guide is designed to help you create your first LLM key on Aptible. The Aptible AI Gateway is a single API to access Anthropic, OpenAI, and Amazon Bedrock, with security and compliance guardrails built in (e.g., encryption, audit logging), and an easy-to-use interface to manage keys, control model access, and use LLMs securely. LLM keys are how you authenticate requests to the AI Gateway, a managed proxy that gives you secure, auditable access to leading AI model providers without needing to manage provider credentials directly in your apps. For a full overview of the AI Gateway, see the AI Gateway documentation.

Create an LLM Key

New Customers

Start your trial to try for free

AI Gateway trial enrollment

From there you’ll be redirected to the LLM Keys dashboard. You’ll receive $10 of LLM credits to get you started, which can be used with any models that we support. If you want to upgrade to a paid plan for more credits, contact Aptible Support.

For those processing PHI: You will need to execute a BAA before you can start processing PHI using the AI Gateway. Contact Aptible Support to execute one.

Create an LLM key

Click Create LLM Key on the AI Gateway Dashboard.You’ll start by creating an Environment, which is how keys and resources are organized in Aptible. This will help you apply model access controls and other permissions to your keys.

Then, create your key.

Give it a descriptive name so you can identify its usage later when reviewing costs and request history.

Once that is complete, you will receive your LLM Key and the Endpoint URL. Make sure to copy the key, as you won’t be able to see it again after you close this window.

Connect to the gateway

You will use this single endpoint to connect to the gateway for all models:

        https://llm-gateway-api.aptible.com

Specific configuration requirements will vary by AI tool. Here are instructions for setting up Claude Desktop to use your key:

Enable Developer Mode: Open Claude Desktop and go to Help > Troubleshooting > Enable Developer Mode.
Access Inference Settings: A new “Developer” menu will appear after Claude restarts. Click Developer > Configure third-party inference.
Enter the Gateway base URL: Select “Gateway” as your connection type, and use https://llm-gateway-api.aptible.com.
Set your Gateway API key: paste your Aptible LLM key here, and set the Gateway auth scheme to “bearer”.
Save and Restart: Click Apply locally, then fully quit and restart Claude Desktop to load the new settings. You can also use LLM keys with any OpenAI-compatible client:

        import openai

        client = openai.OpenAI(
            api_key="YOUR_AI_GATEWAY_KEY",
            base_url="https://llm-gateway-api.aptible.com",
        )

        response = client.chat.completions.create(
            model="anthropic/claude-sonnet-4-20250514",
            messages=[
                {"role": "user", "content": "Hello world!"}
            ],
        )

        print(response)

Start using your LLM Keys!

Once that is complete, you’re ready to start using your LLM Keys securely! All your LLM usage will be encrypted and logged for auditability.When you’re ready to upgrade to a paid AI Gateway plan, contact Aptible Support here or via the Aptible Dashboard. Note: this will require having a payment method on file.

Optionally, start using with PHI

To begin using the AI Gateway with PHI, you’ll need to upgrade to a paid plan and execute a BAA with Aptible. Contact Aptible Support here or via the Aptible Dashboard to continue.Once you have completed these steps, you’re ready to start processing PHI! Aptible will ensure your usage is encrypted, logged, and never stored for model training.

Optionally, configure model access

You can see the available models in the key details page; copy your desired model string and use it wherever you use your LLM key. This view shows which models your key has access to.

Model access is configured at the environment level — all keys within a given environment share the same model access policy. To change which models are available, update the model access settings on the environment.

Review costs and optionally, configure limits

View per-key costs for the current billing period, request history for the past week, and token consumption per request.

​Overview

​Create an LLM Key

Overview

Create an LLM Key