Document OCR in 90+ languages using python (100% free and without internet):
For document OCR, we'll use opensource toolkit Surya.
Python toolkit that does:
• OCR in 90+ languages
• Line-level text detection in any language
• Layout analysis (table, image detection)
• Reading order detection
• Outperforms Tesseract, on par with Google Cloud vision
1. Install the Python library
Run the following command from your terminal. You'll need python 3.9+ and PyTorch.
2. Try out the OCR (text detection) boilerplate code
3. Try out the built-in interactive demo with Streamlit.
Install Streamlit and run the following command in your terminal: 'surya_gui'
🌟 Support the opensource project:
8000+ readers receive my AI newsletter everyday and you are not one of them. Join @_unwind_ai to stay on top of the latest AI developments and tools: github.com/VikParuchuri/s… unwindai.substack.com
• • •
Missing some Tweet in this thread? You can try to
force a refresh
JSON prompting is a way of asking an LLM using a clear, structured format (with keys and values) and expecting the response in the same structured style.
Text prompts → inconsistent, messy outputs
JSON prompt → consistent, parseable data
The Problem With Text Prompts
Natural language is strong, but in AI it’s loose.
“Summarize this email” or “give key takeaways” leaves room for guesswork.
You wouldn’t tell a junior: “Make it better. Do what feels right.”