Skip to content
ImageRouter
Docs

Open WebUI native integration

This guide shows how to combine ImageRouter with your favourite Large-Language-Model in Open WebUI so the same chat can both think and draw.

The recipe works with any LLM supported by Open WebUI (GPT-4, Mistral, Gemini, local Ollama models…) and any image model provided by Image Router.

Open WebUI native integration

Setup

  1. Open http://localhost:3000 (or wherever you host Open WebUI) and click your profile picture (bottom left).

  2. Navigate to the Admin Panel, open Settings, and click Images on the left sidebar.

  3. Change the OpenAI Engine endpoint to https://api.imagerouter.io/v1/openai and enter your API key. Get an API key here if you need.

  4. Toggle Image Generation (Experimental) ON. Disable Image Prompt Generation if you wish (recommended).

  5. Set a default model. Any model from this list.

  6. Click Save. (We can ignore steps and image size as they have no effect on images)

Open WebUI native integration setup

  1. Open a New chat and toggle Image before sending the prompt. You can also ask the LLM to generate a prompt, then click the little image icon under the AI response. Enjoy!