Parse resumes with open models on Chalk GPUs
This demo shows how to run open-source OCR and language models on a Chalk GPU, turning resume PDFs into structured JSON. Upload a resume to try it, then explore the code to see how it works.
Compare a document with its extracted fields
Start with Backend engineer, then try the two-column or scanned sample to compare layouts. You can also upload a PDF.
Compare each field on the right with the source document on the left. Hover a field to highlight its source words; open JSON to see the structured output.
Choose a sample and compare each extracted field with the source page. Chalk serves the GPU models and streams structured results as they are generated.
Built on Chalk
Chalk builds the model container, provisions the GPU, and serves the streaming endpoint. Connect your application to turn resume PDFs into structured JSON.
A model ready to use
Deploy the model and its web app together. Chalk provisions the GPU and returns an HTTPS URL.
Streaming model endpoints
Chalk’s GPU endpoints support streaming with server-sent events. Upload a resume and receive structured JSON as it’s generated, so your application can display results as they arrive.
Deploy from Python
Define the container and its model weights in Python. Later builds can reuse cached image layers.
OCR model weights ready at startup
Chalk builds the OCR model weights into the container image. New replicas load them from the image without downloading them again.
Run it on your own
Clone the example, log in to Chalk, and run the deploy script. Chalk packages OCR and vLLM into a single image, provisions a GPU, and gives you a URL for the demo and API.
To extract data from other document types, update the JSON schema in extract.py and redeploy. Define the fields you need for your document shape, and use the same pipeline to extract them.
git clone https://github.com/chalk-ai/resume-ocr
cd resume-ocr
# Or export CHALK_CLIENT_ID and CHALK_CLIENT_SECRET.
chalk login
# Builds the image on top of the official vLLM image, provisions an
# L4, and prints the URL of your own copy of this demo.
./deploy.py --gpu nvidia-l4Turn your documents into data
Applications, invoices, claims or contracts. Talk to an engineer about running document extraction on GPUs in your cloud.