docs-indic-server

git clone https://github.com/dwani-ai/docs-indic-server.git
cd docs-indic-server

python -m venv --system-site-packages venv

source venv/bin/activate

pip install -r requirements.txt
pip install "numpy<2.0"

python src/server/docs_api.py  --host 0.0.0.0 --port 7861

Dependencies

decord

cd
mkdir external
cd external
git clone --recursive https://github.com/dmlc/decord

cd decord

mkdir build && cd build

export CUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda
export CUDACXX=/usr/local/cuda/bin/nvcc

nvcc --version

cmake .. -DUSE_CUDA=1 -DCMAKE_BUILD_TYPE=Release -DFFMPEG_DIR=/usr

cmake .. -DUSE_CUDA=1 -DCMAKE_BUILD_TYPE=Release

make

cd ../python
python3 setup.py install --user
pip install "numpy<2.0"

olmocr

cd ../../
git clone  https://github.com/allenai/olmocr.git

cd olmocr

pip install --upgrade pip setuptools wheel packaging

 // copy ppyproject.toml
pip install -e .
cd ../../dwani_org/gh-200-docs-indic-server/
pip install "numpy<2.0"

ffmpeg

sudo apt update
sudo apt install ffmpeg
ffmpeg -version

sudo apt update
sudo apt install cmake ffmpeg build-essential python3-dev

sudo apt install pkg-config libavcodec-dev libavformat-dev libavutil-dev libswscale-dev libswresample-dev  libavfilter-dev 

sudo apt-get install poppler-utils

Docs-Indic-Server

Overview

Document parser for Indian languages

git clone https://github.com/slabstech/indic-translate-server
cd indic-translate-server
python3.10 -m venv venv
source venv/bin/activate
pip install -r server-requirements.txt
export HF_TOKEN='YOUR-HF-TOKEN'
export HF_HOME=/home/ubuntu/data-dhwani-models
huggingface-cli download ai4bharat/indictrans2-indic-en-dist-200M
huggingface-cli download ai4bharat/indictrans2-en-indic-dist-200M
python src/server/translate_api.py --port 7861 --host 0.0.0.0 --device cuda --use_distilled

-- For kannad pdf we need to add - NotoSans Kannada font from google

https://fonts.google.com/noto/specimen/Noto+Sans+Kannada

https://github.com/googlefonts/noto-fonts

Contributing

We welcome contributions! Please read the CONTRIBUTING.md file for guidelines on how to contribute to this project.

Also you can join the discord group to collaborate

Reference

Citations

```bibtex citation_1.bib @misc{poznanski2025olmocrunlockingtrillionstokens, title={olmOCR: Unlocking Trillions of Tokens in PDFs with Vision Language Models}, author={Jake Poznanski and Jon Borchardt and Jason Dunkelberger and Regan Huff and Daniel Lin and Aman Rangapur and Christopher Wilhelm and Kyle Lo and Luca Soldaini}, year={2025}, eprint={2502.18443}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2502.18443}, }

<!-- 




wget https://github.com/slabstech/docs-indic-server/blob/01e811210d56e655091313c1df8481d11e7640a6/install-script.sh
chmod +x install-script.sh
bash install-script.sh



## Download Qwen VL

```bash download_model.sh
huggingface_cli download google/gemma-3-4b-it

Download Gemma

```bash download_model.sh huggingface_cli download google/gemma-3-4b-it

## Download Pixtral 

```bash download_model.sh
huggingface_cli download mistralai/Pixtral-12B-2409

Download Moondream2

huggingface_cli  vikhyatk/moondream2

Model Size - 4GB

Getting Started - Development

For moondream, libvips system library is required

sudo apt-get update && sudo apt-get install libvips

Evaluating Results

You can evaluate the ASR transcription results using curl commands. Below are examples for Kannada audio samples.

Kannada

```bash kannada_example.sh curl -s -H "content-type: application/json" localhost:7860/v1/audio/speech -d '{"input": "ಉದ್ಯಾನದಲ್ಲಿ ಮಕ್ಕಳ ಆಟವಾಡುತ್ತಿದ್ದಾರೆ ಮತ್ತು ಪಕ್ಷಿಗಳು ಚಿಲಿಪಿಲಿ ಮಾಡುತ್ತಿವೆ."}' -o audio_kannada.mp3

#### Hindi

```bash hindi_example.sh
curl -s -H "content-type: application/json" localhost:7860/v1/audio/speech -d '{"input": "अरे, तुम आज कैसे हो?"}' -o audio_hindi.mp3

Specifying a Different Format

```bash specify_format.sh curl -s -H "content-type: application/json" localhost:7860/v1/audio/speech -d '{"input": "Hey, how are you?", "response_type": "wav"}' -o audio.wav

### For Production (Docker)
- **Prerequisites**: Docker and Docker Compose
- **Steps**:
  1. **Start the server**:
  For GPU
  ```bash
  docker compose -f compose.yaml up -d
  ```
  For CPU only
  ```bash
  docker compose -f cpu-compose.yaml up -d
  ```

#  - vllm serve vikhyatk/moondream2 --trust-remote-code


## Building Docker Image
Build the Docker image locally:
```bash
docker build -t slabstech/docs_indic_server -f Dockerfile .

Run the Docker Image

docker run --gpus all -it --rm -p 7860:7860 slabstech/docs_indic_server

-->