Skip to content
Nappai Documentation

Vertex AI

This component uses Google’s Vertex AI to create text based on your instructions. You can use it to generate summaries, write different kinds of creative content, answer questions, and more. It’s integrated directly into Nappai’s dashboard for easy use.

Relationship with Google Vertex AI

This component connects to Google’s Vertex AI service, which provides advanced language models. These models are trained on massive amounts of data, allowing them to generate high-quality, human-like text. You don’t need to know anything about Vertex AI directly; this component handles all the technical details.

Inputs

  • Credentials: This is where you provide access to your Google Cloud account. You can either upload a JSON file containing your credentials or let Nappai use the credentials already configured in your environment (this is usually set up automatically). If you’re unsure, leave this blank.
  • Model Name: Choose the specific Vertex AI language model you want to use. The default (“gemini-1.5-pro”) is a good starting point.
  • Project (Advanced): This is the ID of your Google Cloud project. Leave this as the default unless you have a specific reason to change it.
  • Location (Advanced): This specifies the geographic location of the Vertex AI service. Leave this as the default (“us-central1”) unless you have a specific reason to change it.
  • Max Output Tokens (Advanced): This sets a limit on the length of the generated text. A higher number means longer text. Leave this as the default unless you need to control the length.
  • Max Retries (Advanced): If there’s a problem connecting to Vertex AI, this sets how many times Nappai will try again. The default is 1.
  • Temperature: This controls how creative the generated text is. A value of 0.0 means very predictable text, while higher values (e.g., 0.7) will produce more creative and unpredictable results. The default is 0.0.
  • Top K (Advanced): This is an advanced setting that affects the randomness of the text generation. Leave this as the default unless you are familiar with this parameter.
  • Top P (Advanced): This is an advanced setting that controls the probability of selecting words. Leave this as the default unless you are familiar with this parameter. The default is 0.95.
  • Verbose (Advanced): If checked, Nappai will show you more detailed information about the process. Leave this unchecked unless you need extra debugging information.

Outputs

This component doesn’t have explicit outputs in the traditional sense. Instead, it generates text directly within the Nappai workflow. The generated text can then be used by other components in your automation flow, such as a “Text Output” component to display the results or a “Summarizer” to further process the text.

Usage Example

Let’s say you want to generate a short product description for a new widget. You would:

  1. Add the “Vertex AI” component to your Nappai workflow.
  2. (Optional) Upload your Google Cloud credentials.
  3. In the “Model Name” input, select the desired model (or leave it at the default).
  4. In a previous component, provide a prompt such as “Write a short, catchy product description for a new widget that automatically cleans your gutters.”
  5. Connect the output of the prompt component to the input of the Vertex AI component.
  6. Add a “Text Output” component after the Vertex AI component to display the generated description.

Templates

[List of templates where the component is used will be populated here. This section requires information not provided in the original prompt.]

  • Explain SQL results: Use this to understand the results of database queries.
  • Summarizer: Use this to shorten long pieces of text.
  • Text Output: Use this to display the generated text from Vertex AI. [This section requires information not provided in the original prompt. More related components can be added here based on the actual connections in the Nappai system.]

Tips and Best Practices

  • Start with the default settings and adjust them as needed.
  • Experiment with different model names and temperature values to find the best results for your needs.
  • For longer text generation, increase the “Max Output Tokens” value.
  • If you encounter errors, check your Google Cloud credentials.

Security Considerations

  • Ensure your Google Cloud credentials are securely stored and managed. Avoid hardcoding credentials directly into your Nappai workflows. Use environment variables or a secure credential management system.
  • Be mindful of the data you feed into the Vertex AI model. Do not include sensitive or confidential information unless absolutely necessary.