Documentation
LOGIN TO APP →

Configuring an AI Endpoint in Requestador: Steps and Best Practices

Before any action can be executed, Requestador must be able to communicate with an AI provider. This is done by creating an AI endpoint.

What is an AI endpoint?

An AI endpoint defines:

  • which provider is used

  • which model is used

  • how Requestador authenticates

  • which default parameters and headers are sent

Step 1: Open the AI Endpoints section

Log in to Requestador and navigate to the AI Endpoints section.

Click Add AI Endpoint.

Documentation image

Step 2: Fill in the basic information

Complete the following fields:

  • Active - enable this option if the endpoint should be available for use.
  • AI Endpoint Name - enter a clear name for the endpoint, for example:

    • OpenAI GPT-4o

    • OpenAI Product Enrichment

    • Translation Endpoint

  • Description - add an optional description explaining the purpose of the endpoint.
Documentation image

Step 3: Select the API provider

Choose the provider type, for example:

  • OpenAI

Use the provider supported by your Requestador installation.

Step 4: Select the mode

Choose the model that should be used for this endpoint.

Make sure the selected model is available under your provider account and subscription plan.

Step 5: Configure authentication

Choose the authentication type and enter the required credentials.

Example:

  • Authentication type: Bearer Token

  • Value: your API key

Step 6: Add default parameters

If required, add default parameters that should be sent with every request.

Examples may include:

  • temperature

  • max tokens

Step 7: Add optional headers

If your provider requires additional HTTP headers, add them in the Headers section.

Step 8: Test the AI endpoint

Before proceeding to action creation, verify that the AI endpoint is configured correctly and can successfully communicate with the selected provider.

This step is important to confirm that:

  • authentication is valid

  • the selected model is accessible

  • required headers and parameters are correct

  • Requestador can successfully send requests to the provider

If endpoint testing is available in your Requestador environment, use it to validate the connection before saving or using the endpoint in actions.

Step 9: Save the endpoint

Click Save.

The endpoint is now available for use in Requestador actions.

Best practices

  • Use descriptive endpoint names

  • Create separate endpoints for different use cases if needed

  • Validate provider credentials before building actions

  • Keep authentication credentials secure