Dockerfile

FROM python:3.9

# Install system dependencies for GGUF models
RUN apt-get update && apt-get install -y build-essential cmake

# Create cache directory
RUN mkdir -p /tmp/cache && chmod 777 /tmp/cache
ENV HF_HOME=/tmp/cache

WORKDIR /app
COPY . .

# Install Python dependencies
RUN pip install --no-cache-dir -r requirements.txt

# Expose correct port
EXPOSE 8081

CMD ["python", "api.py"]