README_vllm.md

vLLM: local server and integration

Install and run

pip install vllm
python -m vllm.entrypoints.openai.api_server \
  --model /path/to/dots-ocr-model \
  --dtype auto \
  --port 8000 \
  --trust-remote-code

Environment variables

• DOTS_OCR_BASE_URL=http://localhost:8000/v1
• DOTS_OCR_MODEL_NAME=/model
• DOTS_OCR_API_KEY=dummy # vLLM usually does not validate key

Smoke test

curl http://localhost:8000/v1/models | jq

Using with Documentor

OCR config will read the variables from .env via documentor/core/load_env.py automatically.

Docker & Compose

• Dockerfile: Dockerfile.dotsocr
• Docker Compose: compose.yml

Entrypoint script

• Entrypoint script: entrypoint.sh

Models installation

You can install models from Hugging Face.

DotsOCR