-
Updated OpenAI with newer models
-
Added Google Gemini
-
Added AI Horde
-
Added Presets
Note: Only image models that have API keys in your .env file will appear in the dropdown menu
-
Since DALL·E 2 and DALL·E 3 are being deprecated, newer models were added
-
You’ll need a paid OpenAI API key to use these models (Register here: OpenAI)
-
You’ll need a Gemini API key to use these models
-
The free tier still requires a credit card (Register here: Google Gemini)
-
The models in the current code are pay-as-you-go (just like the Open AI models)
-
Create a project and a billing account, then link billing to the project — otherwise the API will return a 429 error
-
A free image API — the tradeoff is slower speeds and older image models (Register here: AI Horde)
-
For the fastest results, pick "first available" model. The results will vary depending on what model is assigned
-
Most models are Stable Diffusion–based
-
AlbedoBase XL (SDXL) typically looks better, but often has high load. The dropdown shows current availability & current load; if load is too high, the quality upgrade is disabled. It can often take 15 minutes to create an image. Open terminal veiw to see updates.
-
To use the AI Horde random function, you’ll also need a free Groq / Llama API key
Install the plugin using the InkyPi CLI, providing the plugin ID & GitHub repository URL:
inkypi plugin install ai_image_multiverse https://github.com/doowylloh88/InkyPi-Image-Multiverse- This plug-in was 100% created using vibe- coding & a lot of yelling at ChatGPT. An actual coder should take over the project to maintain it
- Huge Caveat: I'm in developer mode ONLY. I don't have an e-ink screen yet. They are sold out everywhere. As I wait for stocks to resupply, I'm creating plug-ins I want. I'm running off a Raspberry Pi 5. Most of the image processing happens on the Image Model side, so it shouldn't slam a Pi Zero 2W any more than the existing image plug-in
This project is licensed under the GNU public License.