Ollama-Action

A GitHub composite action that installs Ollama, pulls a model, and runs a prompt — all within your workflow.

Usage

- name: Run Ollama
  uses: MandalAutomations/Ollama-Action@main
  with:
    model: 'tinyllama'
    prompt: 'Explain GitHub Actions in one sentence.'

Inputs

Input	Description	Required	Default
model	The Ollama model to use (e.g. tinyllama, llama3, mistral)	No	tinyllama
prompt	The prompt to send to the model	No	—
cache	Whether to cache the model between runs (true/false)	No	false

Outputs

Output	Description
response	The response from the model

Example Workflow

name: Ask Ollama

on:
  workflow_dispatch:

jobs:
  ask:
    runs-on: ubuntu-latest
    steps:
      - name: Run Ollama prompt
        id: ollama
        uses: MandalAutomations/Ollama-Action@main
        with:
          model: 'tinyllama'
          prompt: 'What is the capital of France?'

      - name: Print response
        run: echo "${{ steps.ollama.outputs.response }}"