Quick Start Guide

Install

• Open the Adobe Creative Cloud Desktop app and install the plugin (double click on .ccx file you downloaded). Open Photoshop and then find Maika in the Plugins menu.
• Make sure you have an API key for the provider you want to use (Replicate, Fal.ai and Google AI Studio).

1. LICENSE tab (activate your license)

• Open the LICENSE tab and paste your license key into the License Key field. Use the « button if you want to reveal the key.
• Click Activate. The status badge should change from "NOT ACTIVATED" to an activated state after validation.

2. SETUP & SETTINGS tab (provider and keys)

API provider and API key setup

• Open the SETUP & SETTINGS tab and choose your API provider from the dropdown list (Replicate, Fal.ai and Google AI Studio).
• Paste your provider API key into the API key field and click Test API key to verify it works. Use « button if you want to reveal the key.
• Enable Save API key if you want Maika to remember the key on this computer (otherwise you will need to paste it again later).

3. GENERATE tab (create or edit images)

Model and number of output images

• Open the GENERATE tab and select the Model you want to use from the dropdown list. Then choose the number of output images (1–3). Start with 1 for testing, or use 2 or 3 to iterate.
• Some Provider/Model combinations may expose additional options.

Prompt

• Write a prompt in the Prompt box. Use natural language or a structured JSON prompt.

Image generation vs Image editing

• Select Image generation to create a new image. Select Image editing when you want to modify a limited area of an existing image while keeping the original composition unchanged. In this mode aspect ratio is disabled.

Input image: how @sel works

• Use a marquee selection to choose what part of the current document will be used as the input image. If no marquee selection is active, the whole canvas will be used. Refer to the input image using the tag @sel in your prompt. Depending on the selected mode @sel represents:
  • Visible layers: current document’s visible composite as input.
  • Active layer: only the currently selected layer as input.
  • None: no input image from current document will be used (in this case do not use the tag @sel in your prompt).
    This mode is only available when Image generation is selected.

Aspect ratio

• Only available for Image generation; it is disabled for Image editing. Select from the dropdown list the aspect ratio that will be used on the output images.

Reference images: how @refN works

• Upload up to 6 reference images using the slots, starting with the one marked + .
• Remove or replace a reference image from its slot using the icons on the thumbnail.
• Rearrange reference images using the drag-and-drop feature.
• Use @ref1, @ref2, … @ref6 in your prompt to refer to each slot.

Generate (and progress)

• Click Generate to start. A progress bar appears below the button while the job runs.
• While the job runs the button changes to Cancel, click it to abort the generation process.

Notes about Google Search / Image search (when available)

• Enable Google Search / Image search appear only with specific provider/model combinations and can help with complex prompts that may require additional search context.

4. SETUP & SETTINGS tab (options)

Input options

• Leave temporary input image open: keeps the temporary input document open (useful for troubleshooting).
• Clear reference slots after generation: automatically clears @ref slots after a successful run.

Output options

• Group output layers: places results into a dedicated layer group for a cleaner layer stack. A new layer group is created for each combination of date - provider - model.
• Add layer mask to output images: when Image editing mode is selected, adds a mask to each imported output layer.
• Fit output images to canvas size: when Image generation mode is selected, scales imported results to fit the current document canvas.
• Assign sRGB color profile to output images: assigns standard sRGB IEC61966-2.1 color profile to imported outputs for consistent color management.

Other settings

• Output layer naming: choose how output layers are named (model/provider/id, model/id, provider/id, or timestamp/id).
• Cancel request if no output after: auto-cancels a request if it takes longer than the time selected from the dropdown list. Select the "Never" option to disable this feature.