Languages Supported

Nanonets supports character/word recognition in over 40+ languages including but not limited to:

1. Arabic
2. Bulgarian
3. Chinese
4. Croatian
5. Czech
6. Danish
7. Dutch
8. English
9. Estonian
10. Filipino
11. Finnish
12. French
13. German
14. Greek
15. Hebrew
16. Hindi
17. Hungarian
18. Icelandic
19. Indonesian
20. Italian
21. Japanese
22. Korean
23. Latvian
24. Lithuanian
25. Macedonian
26. Malay
27. Norwegian
28. Persian
29. Polish
30. Portuguese
31. Romanian
32. Russian
33. Serbian
34. Serbian
35. Slovak
36. Slovenian
37. Spanish
38. Swedish
39. Thai
40. Turkish
41. Ukrainian
42. Vietnamese

Can a single model support multiple different languages?

Yes. You can build a single model to handle documents in multiple languages. It contributes to better accuracy and is easy to deploy.

Is handwritten text supported?

Yes, we support handwritten documents — with some important notes:

• Handwriting must be legible to the human eye. If a human reader struggles to understand the handwriting, AI models may also have difficulty extracting text reliably
• Results may vary depending on writing style, image clarity, and document structure
• We recommend testing a few samples to evaluate performance for your specific use case

Tip: Neat, well-scanned handwritten forms (e.g., block letters or clearly written fields) tend to produce better results