Dockerfile

# ============================================
# 🚀 FINAL OPTIMIZED DOCKERFILE FOR HUGGING FACE SPACES
#    - Sub‑20 second responses on free CPU tier
#    - Multi‑stage build for minimal image size
#    - Open WebUI v0.9.1 (latest stable)
#    - Ollama CPU optimizations baked in
# ============================================

# --- Stage 1: Builder (Install Ollama) ---
FROM python:3.11-slim-bookworm AS builder

# Install Ollama (requires curl + zstd for extraction)
RUN apt-get update && apt-get install -y --no-install-recommends curl zstd \
    && curl -fsSL https://ollama.com/install.sh | sh \
    && apt-get clean && rm -rf /var/lib/apt/lists/*

# --- Stage 2: Final Image ---
FROM python:3.11-slim-bookworm

# Copy Ollama binaries from the builder stage
COPY --from=builder /usr/local/bin/ollama /usr/local/bin/ollama
COPY --from=builder /usr/local/lib/ollama /usr/local/lib/ollama

# Install only essential runtime dependencies
RUN apt-get update && apt-get install -y --no-install-recommends \
    curl ca-certificates \
    && apt-get clean && rm -rf /var/lib/apt/lists/*

# Install the latest Open WebUI (pinned to 0.9.1 for reproducibility)
RUN pip install --no-cache-dir open-webui==0.9.1

# --- CPU & Ollama Optimizations ---
# These environment variables are critical for fast CPU inference
ENV OLLAMA_NUM_THREADS=2 \
    OLLAMA_NUM_PARALLEL=1 \
    OLLAMA_MAX_LOADED_MODELS=1 \
    OLLAMA_KEEP_ALIVE=-1 \
    OLLAMA_CONTEXT_LENGTH=2048 \
    OLLAMA_DEBUG=0 \
    OLLAMA_LLM_LIBRARY=cpu \
    OLLAMA_BASE_URL=http://127.0.0.1:11434

# Disable Open WebUI’s resource‑heavy background features
ENV ENABLE_TITLE_GENERATION=false \
    ENABLE_TAG_GENERATION=false \
    ENABLE_AUTOCOMPLETE_GENERATION=false \
    ENABLE_RAG_WEB_SEARCH=false \
    ENABLE_QUERIES_CACHE=true \
    ENABLE_BASE_MODELS_CACHE=true \
    ENABLE_PERSISTENT_CONFIG=false

# --- Startup Script ---
# Starts Ollama, waits for it to be ready, then launches Open WebUI
RUN echo '#!/bin/bash\n\
set -e\n\
echo "Starting Ollama..."\n\
ollama serve &\n\
OLLAMA_PID=$!\n\
echo "Waiting for Ollama to be ready..."\n\
until curl -s http://127.0.0.1:11434/api/tags > /dev/null 2>&1; do\n\
    sleep 1\n\
done\n\
echo "Ollama is ready."\n\
echo "Starting Open WebUI..."\n\
exec open-webui serve --host 0.0.0.0 --port 7860\n\
' > /start.sh && chmod +x /start.sh

# Hugging Face Spaces expects the app on port 7860
ENV PORT=7860
EXPOSE 7860

CMD ["/start.sh"]