Workflow actions

Generate text with Anthropic

This Workflow Companion action allows you to generate text using Anthropic's Claude AI models. The action can be used with any trigger.

Prerequisites

An Anthropic API key is required to use this action. You can get it from the Claude Console.

Configuration

Basic Setup

  1. Open the app and go to Settings → Actions
  2. Go to the Anthropic (Claude) section
  3. Enter your API key
  4. Set up completion parameters:
    • Name: an identifier for the parameter set.
    • Model: select the Claude model to use. See the models overview for details.
    • Temperature: controls the randomness of the output. Lower values (closer to 0.0) give more focused, deterministic responses; higher values (closer to 1.0) give more creative ones.
    • Max Tokens: maximum length of the generated response. The model may stop before reaching this limit. Check the models overview for model-specific limits.
    • System Prompt: provides context and instructions to Claude, helping define its role or specific goals. See the prompt engineering guide for best practices.
    • Stop Sequences: custom text sequences that will cause the model to stop generating. Enter one sequence per line.
    • Thinking budget or Reasoning effort: how much the model reasons before answering. The field shown depends on the selected model — the Claude 4.1–4.6 budget-thinking models take a token Thinking budget (which must be less than Max Tokens), while newer models take a Reasoning effort level.

Action Parameters

When configuring the action in your workflow, you can specify:

  • Prompt (required): enter a prompt to generate the text that you want. You can use workflow variables in your prompt using the {{variable}} syntax.
  • Image URL: URLs of images, one per line (up to 10). Each image is passed to Claude along with the prompt, so the model can describe, classify, or extract text from them.
  • Completion Parameters: one of the parameter sets you created in the previous step. If not specified, the default parameter set will be used.

Output

The generated text is available as a variable that can be used in subsequent workflow steps.