Quick Start

Get up and running with Simili Bot in 2 minutes

Quick Start

Simili Bot v0.2.0 introduces a Zero-Config path. You can now get semantic duplicate detection and AI-powered triage running entirely within GitHub Actions without setting up an external vector database.

1. Get your Gemini API Key

Simili Bot uses Google Gemini for analyzing issues.

1. Go to Google AI Studio.
2. Create a new API key.
3. Add it to your GitHub repository secrets as GEMINI_API_KEY.

2. Create the workflow

Create a file at .github/workflows/simili-bot.yml in your repository:

name: Simili Bot
on:
  issues:
    types: [opened, edited]

jobs:
  triage:
    runs-on: ubuntu-latest
    permissions:
      issues: write
      contents: read
    steps:
      - uses: similigh/simili-bot@v0.2.0
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          GEMINI_API_KEY: ${{ secrets.GEMINI_API_KEY }}

3. Configuration (Optional)

By default, Simili Bot uses the GitHub Native search backend when no Qdrant URL is provided. To customize the behavior, create a simili.yaml file in your repository root:

search:
  backend: "github_native" # Powered by GitHub's hybrid search API
  bm25_fallback: true      # Fallback to local keyword search if API fails

defaults:
  similarity_threshold: 0.70
  max_similar_to_show: 3

Done! 🎉

Next time an issue is opened or edited, Simili Bot will automatically:

1. Search for similar issues using GitHub’s native hybrid search.
2. Analyze potential duplicates using Gemini.
3. Suggest labels and provide a quality assessment.
4. Post a summary comment to the issue.

---

Next Steps

Advanced Setup

Scale to multiple repositories or use Qdrant for advanced semantic search.

Claude Code

Automate your pull request workflows and document synchronization.