Gemini Web Chat

A minimal no-login chat web app:

• users open the page and start chatting immediately
• the browser sends requests to your server
• the server forwards them to Gemini with GEMINI_API_KEY
• the API key stays on the server

Local run

python -m venv .venv
.venv\Scripts\Activate.ps1
pip install -r requirements.txt

Create .env in the project root:

GEMINI_API_KEY=your_real_key
GEMINI_MODEL=gemini-2.5-flash
GEMINI_MODELS=gemini-2.5-flash,gemini-3-flash-preview,gemini-2.5-flash-lite,gemini-3.1-flash-lite-preview
PORT=8000
GEMINI_CONNECT_TIMEOUT=10
GEMINI_READ_TIMEOUT=180

Then run:

python app.py

Open http://127.0.0.1:8000

Notes

• default model is gemini-2.5-flash
• GEMINI_MODEL can override the default
• GEMINI_MODELS controls fallback order when a model hits quota or rate limits
• the app loads .env automatically on startup
• long Gemini responses can use GEMINI_CONNECT_TIMEOUT and GEMINI_READ_TIMEOUT
• chat history lives in browser memory and clears on refresh

Ubuntu 22.04 deployment

Upload the project to the server, then run:

chmod +x deploy_ubuntu22.sh
./deploy_ubuntu22.sh

The script will:

• switch Ubuntu 22.04 apt sources to Tsinghua Tuna by default
• install Python packages needed on Ubuntu 22.04
• copy the project to /opt/gemini-web-chat
• create a virtual environment
• install dependencies with Tsinghua PyPI by default
• generate a systemd service

If you do not want the script to replace apt sources, run:

USE_TUNA_APT=0 ./deploy_ubuntu22.sh

If you do not want to use Tsinghua PyPI, run:

USE_TUNA_PIP=0 ./deploy_ubuntu22.sh

After that:

sudo nano /opt/gemini-web-chat/.env
sudo systemctl enable --now gemini-web-chat
sudo journalctl -u gemini-web-chat -f

Open:

http://your-server-ip:8000