How to generate images in bulk from a prompt list
Generating one image at a time does not scale to a video that needs dozens of visuals. This guide shows the workflow in Melia: one text file of prompts in, one folder of numbered images out.
1. Write the prompt list
A prompt list is a plain text (.txt) file. Put one prompt per block and separate blocks with a line that contains only three dashes. Numbering each prompt (1. 2. 3.) works too.
A cozy wooden cabin beside a frozen lake under the northern lights, cinematic lighting, 16:9 --- A giant ancient turtle carrying a village on its shell, fantasy world, sunrise, 16:9 --- An underwater research laboratory surrounded by glowing jellyfish, futuristic, cinematic, 16:9
You can also write and edit prompts inside Melia's Prompt Editor. Melia does not write the prompts for you: you decide the shots, Melia handles the volume.
2. Choose files, provider and model
In Melia choose the prompts file and an output folder, then pick Gemini or OpenAI and a model. You add your own API key once in Settings; generation is billed by your provider, not by Melia. Start with two or three prompts to check the style and the cost before running a large list.
3. Run the batch
Press Generate. The dashboard shows live progress, completed, remaining and failed counts, and a preview gallery. You can raise the number of parallel workers in the run settings; failed images are retried automatically.
4. Stop and resume safely
Stop keeps every finished image. If you close the app or the run is interrupted, press Generate again and choose Resume: Melia continues in the same run folder and only creates the images that are still missing.
5. Collect the numbered output
Each run gets its own folder, and images are named in prompt order:
Run001/ 0001.png ← prompt 1 0002.png ← prompt 2 0003.png ← prompt 3
The Gallery shows everything in the output folder and Reports lists each run with its image count and estimated cost (an estimate — your provider's invoice is the real figure). Drop the numbered files into your editor and they stay in story order.
Tips for consistent results
- Repeat a short style phrase in every prompt (for example the lighting, lens and aspect ratio) so the batch looks like one video.
- Keep prompts self-contained: each image is generated independently.
- Run a small test batch first, then the full list.
Next: set up Gemini, set up OpenAI or read the getting started guide.