Getting Started

Getting Started

Get a working Prompt2PR workflow in under 5 minutes.

---

Prerequisites

• A GitHub repository where you have admin access
• An API key from one of the supported LLM providers (or a GitHub Copilot subscription for GitHub Models)

Step 1: Get an API Key

Choose a provider and get your API key:

Provider	Where to get the key
Mistral	console.mistral.ai → API Keys
OpenAI	platform.openai.com → API Keys
Anthropic	console.anthropic.com → API Keys
GitHub Models	No key needed — uses the built-in GITHUB_TOKEN

Step 2: Add the Secret

In your GitHub repository:

1. Go to Settings → Secrets and variables → Actions
2. Click New repository secret
3. Add your API key with the correct name:

Provider	Secret Name
Mistral	MISTRAL_API_KEY
OpenAI	OPENAI_API_KEY
Anthropic	ANTHROPIC_API_KEY
GitHub Models	No secret needed

Step 3: Create a Workflow File

Create .github/workflows/prompt2pr.yml in your repository:

name: Prompt2PR
on:
  workflow_dispatch:
    inputs:
      prompt:
        description: 'What should the AI fix?'
        required: true

permissions:
  contents: write
  pull-requests: write

jobs:
  prompt2pr:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: davd-gzl/Prompt2PR@v1
        with:
          prompt: $
          provider: mistral
        env:
          MISTRAL_API_KEY: $
          GITHUB_TOKEN: $

Note: Replace provider: mistral and MISTRAL_API_KEY with your chosen provider and its corresponding secret. {: .notice–info }

Step 4: Enable PR Creation

Go to Settings → Actions → General → Workflow permissions and check:

• “Allow GitHub Actions to create and approve pull requests”

Without this, the action will fail with a permission error.

Step 5: Trigger the Workflow

1. Go to the Actions tab in your repository
2. Select the Prompt2PR workflow
3. Click Run workflow
4. Enter your prompt (e.g., “Add missing error handling to all async functions”)
5. Click Run workflow

The action will scan your files, call the LLM, apply the changes, and open a PR. Check the workflow run logs for details on what was changed.

---

What Happens Next

After the workflow runs successfully:

• A new branch is created (e.g., prompt2pr/1707900000)
• A Pull Request is opened with the LLM’s changes
• The PR body includes the original prompt, a summary of changes, and metadata
• The prompt2pr label is applied

Review the PR like any other — merge it, request changes, or close it.

---

Using GitHub Models (No API Key)

If you prefer not to manage external API keys, GitHub Models lets you use LLMs directly through GitHub’s infrastructure:

permissions:
  contents: write
  pull-requests: write
  models: read # Required for GitHub Models

jobs:
  prompt2pr:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: davd-gzl/Prompt2PR@v1
        with:
          prompt: 'Fix typos in documentation'
          provider: github
          model: openai/gpt-4o
        env:
          GITHUB_TOKEN: $

Important: GitHub Models requires the models: read permission and a GitHub Copilot subscription. {: .notice–warning }

---

Next: Configuration Reference