Docker Deployment

Quick Start

# Pull the image
docker pull ghcr.io/raphaelmansuy/edgeparse:latest

# Process a PDF
docker run --rm -v $(pwd):/data ghcr.io/raphaelmansuy/edgeparse:latest \
  edgeparse /data/document.pdf -f markdown

Dockerfile

FROM rust:1.83-slim AS builder
WORKDIR /app
COPY . .
RUN cargo build --release --bin edgeparse-cli

FROM debian:bookworm-slim
COPY --from=builder /app/target/release/edgeparse-cli /usr/local/bin/edgeparse
ENTRYPOINT ["edgeparse"]

Docker Compose

services:
  edgeparse:
    image: ghcr.io/raphaelmansuy/edgeparse:latest
    volumes:
      - ./input:/data/input
      - ./output:/data/output
    command: /data/input/document.pdf -f json --output-dir /data/output

Image Size

The production image is approximately 15 MB — no Python runtime, no Java, no ML models.