Chat with your repo in under 2 minutes using GitHub Actions on supported Cloud providers.

Repochat is a LLM chatbot with stateless data ingestion capabilities through its API.

• A supported cloud provider credentials ;
• An OpenAI API key.

jobs:
    steps:
    - name: Deploy a convenient chatbot for your repo
      id: deploy_repochat
      uses: flavienbwk/repochat-action@v0
      if: github.ref == 'refs/heads/main'
      with:
        dirs_to_scan: "./example,README.md"  # comma-separated glob dirs to analyze from this repo (required)
        openai_api_key: ${{ secrets.OPENAI_API_KEY }}  # (required)
        openai_model_type_inference: "gpt-4o-mini"  # (required)
        openai_model_type_embedding : "text-embedding-3-small"  # (required)
        cloud_provider: 'scaleway'  # (required)
        provider_key_id: ${{ secrets.PROVIDER_KEY_ID }}
        provider_key_secret: ${{ secrets.PROVIDER_KEY_SECRET }}
        provider_project_id: ${{ secrets.PROVIDER_PROJECT_ID }}
        provider_default_region: 'fr-par'
        provider_default_zone: 'fr-par-2'

    - name: Get container domain from precedent step
      run: echo "DOMAIN=${{ steps.deploy_repochat.outputs.domain }}" >> $GITHUB_OUTPUT
      id: repochat_domain

• Scaleway
  • Refer to Scaleway's documentation to generate API keys.
  • Additional required parameters:
    • cloud_provider: 'scaleway'
    • provider_key_id: ${{ secrets.PROVIDER_KEY_ID }}
    • provider_key_secret: ${{ secrets.PROVIDER_KEY_SECRET }}
    • provider_project_id: ${{ secrets.PROVIDER_PROJECT_ID }}
    • provider_default_region: 'fr-par' # example
    • provider_default_zone: 'fr-par-2' # example

make dev

1. Increase repochat's version in ./package.json

2. Run build and commit latest edits:

   npm run build
   # git add && git commit && git push...

3. Merge on main

   This will create a release based on package.json and push the :latest Docker image.

• Directory data chat
• API data chat
• GitHub Actions release
• Secure ingestion endpoint (/api/ingest)
• Secure RepoChat with optional password

Vercel is very limited when deploying everything but JS. First, ChromaDB (and any sqlite-based library) is not supported in Vercel. Then, this project uses a FastAPI Python API that requires more storage than Vercel's 250MB bundle limit.