Guide/Evoke
Image Generation

Evoke

Generate stock-ready images using Google's Imagen — one of the most photorealistic AI image models available. Paste prompts, hit generate, get images.

What It Does

Evoke connects to Google Imagen to generate high-resolution images suitable for stock platforms. It handles authentication, queuing, and output automatically. Your job is to bring the prompts — Evoke does everything else.

Getting Started

You need a Google Cloud project with Vertex AI enabled and a Service Account key. It sounds technical but takes about 10 minutes to set up:

1. Create a project at console.cloud.google.com

2. Enable the Vertex AI API

3. Create a Service Account and download the JSON key file to your computer

4. In Evoke Settings, click Browse and select the downloaded JSON file

The file path is saved automatically — authentication renews in the background and you won't need to touch it again.

Generating Images

Paste one prompt or import an entire approved batch from Rune. Choose your aspect ratio — 1:1 for social and editorial, 3:2 or 16:9 for landscapes, 2:3 for portrait subjects. Hit Generate.

Up to 4 images generate at the same time, so large batches move fast. A batch of 100 prompts typically finishes in 20–30 minutes depending on your connection.

Live Queue

Every job is visible in the queue in real time — queued, generating, completed, or failed. Failed jobs show the error message so you know whether to retry with the same prompt or adjust it first. Completed images show a thumbnail so you can spot obvious failures without opening the folder.

Output & Next Steps

Images save to your chosen output folder with timestamps in the filename, so batches stay sorted automatically. From here, most workflows continue into Obscure (for editing — background removal, upscaling, relighting) and then Grimoire (for metadata generation before uploading to Adobe Stock or Shutterstock).