Credentials are a way to store a model provider’s authentication details in order to use them when connecting a model to the team. You can create and manage Credentials from the corresponding tab in the Agent Hub.

Supported providers

Provider	Required parameters	Optional parameters
OpenAI	API Key	Organization ID - A unique organization ID, which ensures requests are correctly attributed to the right subscription.
Anthropic	API Key	–
Gemini	API Key	–
Fireworks AI	API Key	Base URL - For Direct-Route deployments, set your direct deployment URL.
Vertex AI	Service Account Key (JSON) - A key with access to Vertex AI in your GCP service account. Region - Enter the region of the Vertex AI resource region, for example `us-central1` or `global`.	Project - Enter a Vertex project ID if the project is different from the one in the Service Account Key.
Databricks	API Key - This is your access token. Base URL - Provide an API base URL pointing to the instance’s model serving endpoints.	–
AWS Bedrock	API Key - Long-term Bedrock API key. Region - Enter the region of the AWS Bedrock resource region, for example `us-east-1`.	–
Custom	Base URL - This is your selected model's API endpoint URL.	API Key

Create credentials

To create Credentials:

From the left panel, go to the Agent Hub.
In the Credentials tab, click + Create.
Type the name of your credentials.
Type the description of your credentials (optional).
Select a provider from the list. For Custom models, continue from here.
Enter your provider’s API Key. For safety reasons, the API key will not be visible after creating the credentials.
Enter the Organization ID, if the selected provider’s authentication details are tied to a specific one (OpenAI only).
Click Create.

After your Credentials have been created, a call will be made to the provider to ensure the API Key is valid and active, and that the Organization ID matches it. Once verified, you can use them normally in any model connection, provided the credentials correspond to the provider of the model you’re trying to connect.

📘
Keep in mind
You can only have up to 200 Credentials per team.

Setting up custom model credentials

You can connect your custom AI models or on-premises deployments to SuperAnnotate by choosing Custom Models in Agent Hub. This feature is ideal for organizations that:

Have on-premises AI infrastructure
Use AI providers not natively supported by SuperAnnotate
Want to maintain complete control over their models and data
Need to comply with specific security or regulatory requirements

📘
Key benefits
Keep your models internal, maintain security control, use any AI provider, and integrate seamlessly with SuperAnnotate's orchestration and monitoring tools.

Requirements: what you need to provide

Requirement	Status	Description
API Endpoint URL	Required	The base URL of your model's API endpoint (must be HTTPS).
API Key	Optional	Authentication credentials if your endpoint requires them.
OpenAI Format API	Required	Your endpoint must implement the OpenAI Chat Completions API format.

🚧
Critical Requirement
Your API endpoint must follow the OpenAI Chat Completions API format. This is non-negotiable for ensuring compatibility with SuperAnnotate's platform.

Technical Prerequisites

HTTPS endpoint: Your API must be accessible via HTTPS (not HTTP).
Public accessibility: The endpoint must be reachable from the internet.

Setup Guide

Step 1: Prepare Your API Endpoint

Ensure your API gateway implements the OpenAI Chat Completions format. Your endpoint should:

Accept POST requests to /v1/chat/completions
Process requests in OpenAI format (see API Format below)
Return responses in OpenAI format

Step 2: Configure Network Access

Optionally whitelist SuperAnnotate's infrastructure:

IP: List of SuperAnnotate IPs you may need to whitelist
Domain: http://hooks.superannotate.com/

Note: Whitelisting is optional but recommended for additional security control.

Step 3: Create Credentials in SuperAnnotate

From the Agent Hub, go to the Credentials tab.
Click + Create Credentials.
Select Custom from the Provider dropdown.
Fill in the required fields:
- Base URL: Your API endpoint (e.g., https://api.yourcompany.com/v1)
- API Key: Your authentication key (optional)
Click Save.

Step 4: Connect Your Model

From the Agent Hub, go to the Models tab.
Click + Connect Model.
Select the credential you just created in the previous step.
Provide your model’s name or ID.
Configure any custom parameters if needed.
Click Connect.

Edit credentials

To edit your credentials:

In the Credentials tab, find the Credentials you want to edit.
Click on the three dots ⋮, and then click Edit.
Make your changes accordingly. You can edit everything except for the Provider.
Click Save.

Any changes made to the authentication details will require another call to be made to the provider to verify the connection again. Once it is verified, you may use the credentials again.

🚧
Editing your credentials
If there are any issues with the connection’s credentials (i.e. your tokens were exhausted), editing will not take effect.

Delete credentials

In the Credentials tab, find the Credentials you want to edit.
Click on the three dots ⋮, and then click Delete.
Click Delete.

📘
Keep in mind
When you delete credentials, any models using them will become invalid.