Ollama support

[joshendriks]

Prerequisites

• [x] I have checked the latest documentation and this feature doesn't exist
• [x] I have searched for similar feature requests and found none
• [x] I have linked to related issues or discussions in the description (if any exist)

Type of Improvement

New Functionality

Proposed Solution

I would like to have a Ollama Invoker

My request is straight forward, just like the other model configurations ollama support would allow user to configure somtehing like:

{
  "name": "ollama-llama3.3",
  "type": "ollama",
  "model": "llama3.3",
  "url": "http://localhost:11434"
}

Where model is the exact model name as displayed on the ollama website: https://ollama.com/library/llama3.3

The integration could be made so that the model is automatically downloaded. But that is just a wish.

I can help to implement and test.

[0GiS0]

Hi @joshendriks 👋🏻

You can use Ollama with Prompty today if you choose the openai type and use as base_url your Ollama URL server + /v1 as endpoint (for example http://localhost:11434/v1). Here you can find an example in my blog (sorry it's in Spanish but the example is in English 😃): https://www.returngis.net/2025/04/como-usar-prompty-con-ollama/

Hope this helps!

[joshendriks]

Great stuff @0GiS0, thanks a l lot!