Literal AI icon

Literal AI

Use the Literal AI API

Overview

This node integrates with the Literal AI API to manage various resources related to AI projects, including prompts. Specifically, for the "Prompt" resource and the "Get A/B Testing" operation, it retrieves A/B testing configurations associated with a named prompt. This is useful in scenarios where you want to analyze or manage different prompt variants being tested against each other to optimize AI model outputs.

Practical examples include:

  • Fetching current A/B test setups for a prompt to monitor performance.
  • Integrating prompt A/B testing data into workflows for automated decision-making or reporting.

Properties

Name Meaning
Name The name of the prompt for which to retrieve A/B testing information. This is a required string input.

Output

The node outputs a JSON object under the content field containing the A/B testing configuration details for the specified prompt. The exact structure depends on the Literal AI API response but generally includes information about the different prompt variants involved in the A/B test, their rollout percentages, and possibly performance metrics.

No binary data output is produced by this operation.

Example output structure (conceptual):

{
  "content": {
    "promptName": "example-prompt",
    "abTesting": [
      {
        "variantId": "variant1",
        "rollout": 50,
        "metrics": { /* performance data */ }
      },
      {
        "variantId": "variant2",
        "rollout": 50,
        "metrics": { /* performance data */ }
      }
    ]
  }
}

Dependencies

  • Requires an API key credential for the Literal AI service to authenticate requests.
  • The node uses the official Literal AI client library (@literalai/client) bundled within the node.
  • No additional environment variables are explicitly required beyond the API key credential.

Troubleshooting

  • Common issues:

    • Invalid or missing API key credential will cause authentication failures.
    • Providing a prompt name that does not exist or has no A/B testing configured may return empty results or errors.
    • Malformed JSON inputs in other operations can cause parsing errors, but for this operation only a simple string name is required.
  • Error messages:

    • Authentication errors typically indicate invalid or expired API keys; verify and update credentials.
    • Not found errors suggest the prompt name does not match any existing prompt; check spelling and existence.
    • JSON parsing errors are unlikely here since only a string parameter is used, but if encountered elsewhere, ensure JSON inputs are valid.

Links and References

Discussion