Image Generation Actions (Native + DALL·E 3)

Overview

Taskade automations include two image-generation actions: a native Generate Image action (Taskade-managed, no API key needed) and Generate Image with DALL·E 3 (bring-your-own OpenAI key). Both are perfect for generating illustrations, concept art, or any image content inside an automation. Combine with Ask Agent to have an AI write the prompt from a brief, then Upload File to Media to store the result.

The bulk of this guide walks through the DALL·E 3 flow, since it's the action that requires API setup. Jump to Native Generate Image Action for the no-key option.

💡 Note: To use this feature, you'll need an OpenAI API key with billing enabled and a positive balance. Visit https://platform.openai.com/docs/guides/images?api-mode=responses to learn more.

How to Get Started

Add the Action

• Create a new automation flow → ➕ Add Step.
• Choose the Generate Image with DALL·E 3 action.

Obtain OpenAI API Key

• Visit your OpenAI dashboard: https://platform.openai.com/api-keys.
• Sign in to your OpenAI account, or create one if you don’t have one already.
• Click ➕ Create new secret key.

• (1) Label the new key (optional).
• (2) Click Create secret key.

• Copy your key and keep it private.

❗ Important: Save your key in a safe place. You won't be able to view it again.

• Click Done to finish the process.

Authorize the Integration

• Once the action is added, you’ll need to add your API Key.
• In the Header section, replace YOUR_API_KEY with your OpenAI API key.

💡 Note: When using your OpenAI API key, keep the word "Bearer," followed by a single space, then your API key, exactly like this:

Bearer YOUR_API_KEY

• Click Save changes to confirm the changes.

Define the Image Request

• The JSON Body section "tells" the action what to generate. You can configure the prompt yourself or click ➕ to pull the data from actions/triggers.

• model: Always set to dall-e-3.

• prompt: Describe the image you want. You can be creative and detailed — the more specific, the better the results.

• n: Number of images to generate.

• size: The image resolution. Common options include:

• 1024x1024 (square)

• 1024x1792 (portrait)

• 1792x1024 (landscape)

Set Response Schema

After the image is generated, the API will send a response that includes a link to the image. You can then use the generated response link in subsequent steps:

• In the Response Body Type section:

• created: A timestamp of when the image was generated.

• data: (This holds the actual image(s)).

• Each object contains:

• url: (string) — the link to the image.

Native Generate Image Action

Taskade also ships a native Generate Image action that calls a Taskade-managed image model from any flow step and writes the result straight into the Media tab. No API key, no OpenAI account, no billing setup. Available wherever the rest of the AI actions run.

• One step in, one image file out, written to your workspace Media
• Reuse generated images in Taskade Genesis apps and chat threads
• Pairs with the Categorize and Branch actions for routing by image type
• Pick the model on the step itself when you want to compare outputs

When to choose each option

For most users the native action is the right starting point. Drop in the DALL·E 3 action when you have an OpenAI account already wired up or need that exact model.

Use Cases

Not sure how to make the most of this action? Here are a few ideas:

Related guides

• Image Generation in Agents — Native Taskade image generation (no API key)
• Media Upload Action — Save generated images
• Agent Action — AI-generated image prompts
• HTTP Request — Similar pattern for other external APIs