I Built Exocortex MediaForge for Game Asset Batches

Why I made it

Doing image generation one prompt at a time in a browser is fine until you actually need a folder full of usable assets.

Once it turns into real batch work, the annoying parts take over: copying prompts around, downloading files one by one, cleaning up filenames, remembering which settings produced what, and retrying the few items that failed without restarting everything.

I wanted a small local utility that handled that part better.

It started as an image batch tool, and now it supports OpenAI video plus local ComfyUI image and video workflows too.

What it does

Exocortex MediaForge is a Windows desktop app that takes prompt lists and turns them into organized output on disk.

You can:

• paste prompts directly
• load .txt, .csv, or .json prompt files
• switch between OpenAI and local ComfyUI
• switch between image mode and video mode
• use a few curated local ComfyUI presets for SDXL, FLUX, Chroma, Wan, and HunyuanVideo
• choose an output folder
• run the batch locally with your own OPENAI_API_KEY
• keep a manifest.json and generation_log.csv
• retry failed items without rerunning the whole list

The point is not to build a big platform around image generation. It is just a simpler desktop workflow if you already know what you want to generate.

Why I kept it local

This kind of job makes more sense as a local tool.

You already have folders. You already have prompt lists. You probably do not need accounts, dashboards, invites, or telemetry wrapped around the process. You just need the batch to run, the files to land in the right place, and the output to stay organized.

In OpenAI mode, the generation requests still go to OpenAI because that is how the APIs work. In ComfyUI mode, the app talks to your own local server and workflow. In both cases, the surrounding workflow stays local and direct.

Good fit for asset batches

It is especially useful for the kind of work where you want structure more than novelty:

• enemy sets
• prop packs
• pickups and consumables
• icon batches
• environment texture concepts
• prompt-list variations you want saved cleanly
• short generated clips you want saved cleanly as MP4s

If your prompt files already carry categories, asset IDs, or filename prefixes, the app keeps that structure instead of dumping everything into one messy folder.

What gets written

The app saves generated images or MP4 videos directly to the output folder you choose.

Alongside the images, it writes:

• manifest.json with prompt, filename, status, settings, and path data
• generation_log.csv as a flat run log

That makes it easier to inspect a batch later without guessing what happened.

Download

If you want a local batch workflow for image generation, it is here:

• Open the tool page
• Download the Windows ZIP
• Open the GitHub repo
• Open the GitHub release

Use your own OpenAI API key, pick an output folder, and let it handle the boring part.