Minion Chat

A web-based Minion chatbot that speaks in Minionese, powered by Llama (local via Ollama or cloud via Groq). Includes minion-style TTS and a simple web UI.

Run locally

Option A: With Ollama (local Llama)

1. Install Ollama and run: ollama pull llama3.2
2. Install FFmpeg and SoX (for minion voice). On Windows, set paths in env if needed.
3. Install deps and run:

pip install -r requirements.txt
python app.py

4. Open http://127.0.0.1:5000

Option B: With Groq (cloud Llama, no Ollama)

1. Get a free API key at console.groq.com/keys
2. Set it and run:

pip install -r requirements.txt
set GROQ_API_KEY=your_key_here
python app.py

3. Open http://127.0.0.1:5000

---

Deploy as a web app (GitHub + Render)

You can host the app on the web so anyone can chat with the Minion. The app runs in Docker on Render.com (free tier) and uses Groq’s free Llama API (no GPU or Ollama on the server).

1. Push the project to GitHub

1. Create a new repo on GitHub (e.g. minion-chat).
2. In your project folder:

git init
git add app.py index.html requirements.txt Dockerfile .dockerignore render.yaml .env.example README.md
git add minion.png
git commit -m "Minion Chat web app"
git branch -M main
git remote add origin https://github.com/YOUR_USERNAME/minion-chat.git
git push -u origin main

(Replace YOUR_USERNAME and minion-chat with your repo. Add minion.png only if you have one.)

2. Deploy on Render

1. Go to dashboard.render.com and sign in with GitHub.
2. New → Web Service.
3. Connect the minion-chat (or your repo) and confirm:
   • Runtime: Docker
   • Dockerfile path: ./Dockerfile (or leave default)
4. Under Environment, add:
   • Key: GROQ_API_KEY
   • Value: your Groq API key from console.groq.com/keys
5. Click Create Web Service. Render will build the Docker image (installs FFmpeg + SoX) and start the app.
6. When it’s live, open the URL Render gives you (e.g. https://minion-chat.onrender.com).

That URL is your web-based Minion app running on GitHub (code) + Render (hosting) with the Llama model via Groq.

3. Optional: render.yaml (Blueprint)

The repo includes a render.yaml so you can use Blueprint deploy: in Render, New → Blueprint and point it at your repo. You still need to set GROQ_API_KEY in the Render dashboard for the service.

---

Environment variables

Variable	Used when	Description
GROQ_API_KEY	Web / cloud	Groq API key (get at console.groq.com). If set, the app uses Groq instead of Ollama.
GROQ_MODEL	Optional	Groq model (default: llama-3.3-70b-versatile).
OLLAMA_HOST	Local	Ollama URL (default: http://127.0.0.1:11434).
OLLAMA_MODEL	Local	Ollama model (default: llama3.2).
FFMPEG_PATH	Optional	Path to FFmpeg (auto on Linux/Docker).
SOX_PATH	Optional	Path to SoX (auto on Linux/Docker).

Copy .env.example to .env and fill in values if you want to use a local env file.

---

Features

• Llama-powered replies: Local (Ollama) or cloud (Groq), always in Minionese.
• Minion voice: TTS with pitch/tempo effects (gTTS → FFmpeg → SoX).
• Short, in-character answers: Keyword-style Minion replies; fallback phrases if the model sounds too “English”.
• Web UI: Single-page chat; optional name; works in the browser on your machine or on Render.

---

License

Use and modify as you like. Minion/Despicable Me are trademarks of Universal/Illumination; this project is a fan app, not affiliated with them.