Using Gemini with Capture One Actions

Gemini is an AI model designed for open-ended creative workflows. Unlike predefined AI services, Gemini uses detailed prompts and optional reference images to shape and guide the final output.

With Capture One Actions, Gemini is used directly from within the capture workflow. Actions define the base prompt and parameters, while users can refine results by appending text, adding image references, or using annotations. This makes Gemini well suited for creative exploration that benefits from iteration.

Contents

• Setting up the Gemini Connector and Actions
• Example 1: Ghost mannequin composites during production
• Example 2: Creative background exploration with prompts and annotations

Setting up the Gemini Connector and Actions

Setting up Gemini with Capture One Actions follows this process: you start with obtaining a Gemini API key, then configure the Connector and create Actions that define prompts, parameters, and optional user inputs.

Step 1: Configure the Gemini Connector

1. Sign in to your Capture One account and open the web-based admin platform. Studio for Teams users can access it from the Actions tool in Capture One, by selecting Manage Actions.
2. Navigate to the Connectors tab. 
   
3. Choose Gemini and click Connect.
4. Enter your Gemini API key.
5. Save the Connector.

Step 2: Create Gemini Actions

Gemini gives you full creative control through a prompt-driven configuration. Each Action defines a base prompt and optional parameters that guide how the generative model produces results.

1. Select Create Action in the web-based admin platform.
   
2. Enter an action name that clearly describes the intended output.
   
3. Set Service to Gemini.
4. Select the Gemini model that best fits your scenario. Additional image and video models may be added over time.
5. Write the main Prompt that defines the desired output. You can use any LLM to help draft and refine it.

Effective prompts describe how the subject should be presented, including visual style, environment, camera angle, composition, lighting, and mood. It is common to iterate on prompts and refine them over time. Many teams use a large language model to help author and tune prompts before finalizing them.

You can also define parameters that allow further control at Action creation time or directly inside Capture One:

• Image references
  Add up to four image reference parameters. These allow users to send additional images alongside the main variant, such as interior garment details for a ghost mannequin workflow.
• Custom fields
  Add text fields, text areas, number fields, or image fields to extend or customize the prompt. Text areas are commonly used to allow users to append instructions inside Capture One.

If specific parameter names are used, they have special meaning for Gemini:

• mediaResolution (text): LOW, MEDIUM, or HIGH
• temperature (number): controls creativity, from 0 to 2
• seed (number): enables repeatable results
• maxOutputTokens (number): limits response size
• imageConfig.imageSize (text): 1K, 2K, or 4K
• imageConfig.aspectRatio (text): 1:1, 2:3, 3:2, 3:4, 4:3, 9:16, 16:9, or 21:9

Once configured, save the Action in draft mode so it can be tested before being published to the wider team.

Example 1: Ghost mannequin composites during production

Scenario

A studio is shooting apparel on mannequins and wants to generate ghost mannequin composites during the shoot. The goal is to validate garment shape, alignment, and overall presentation before the set is wrapped.

A Gemini Action has been created with a base prompt for ghost mannequin compositing.

How it is done in Capture One

1. Capture the exterior garment image.
   
2. Select the image and apply the Ghost mannequin Gemini Action.
3. Add an interior garment image as an image reference.
4. Trigger the Action from the Actions tool.
5. Review the generated composite when it returns.

Result

Gemini generates a composite image with a hollow, three-dimensional effect. The team can assess fit, alignment, and presentation on set, reducing the risk of reshoots.

Example 2: Creative background exploration with prompts and annotations

Scenario

A creative team wants to explore lifestyle background options for a product without committing to a single look. The goal is to iterate quickly and steer results collaboratively during the shoot.

A Gemini Action has been created with a base background prompt and allows users to append text and add annotations.

How it is done in Capture One

1. Select the images to process.
   

2. Create the Background generation Gemini Action with the imageConfig.aspectRatio parameter set to 16:9 and the following prompt:
   

    

3. Trigger the Action and review the result.

Result

Gemini generates creative background variations guided by both text and visual input. Teams can iterate quickly, compare options, and converge on a creative direction while still on set.