AI Use Cases

Work · by Datalab

Surya

OCR, layout and reading-order models covering 90+ languages, weights included.

+ Beats the old OCR engines on messy scans and non-Latin scripts, locally.
− The licence is not plain open source above a revenue line - read it before shipping.

Open sourceWeights are free for research and small companies; a commercial licence applies above a revenue thresholdopen sourcecliapilinuxmac

Open Surya ↗ best for: scanned archives in languages Tesseract mangles

Evidence

rank #103 of 333
42evidence score / 100
21,542GitHub stars
0Hacker News stories, 30d · query "Surya"
0mentions in our release feed, 30d
4days since last activity
Ranked forExtract data from documents

Head to head: Unstructured vs Surya Docling vs Surya Surya vs LlamaParse

Surya alternatives

All 6 alternatives →
1
Unstructured Open source
Ingest library and hosted pipeline that normalises 25+ file types into chunks for retrieval.

+ The connector list is the real product - S3, SharePoint, Confluence and Drive land as clean chunks.
− The open library's best parsing paths push you toward the paid API.

Apache 2.0 core; the hosted platform and connectors are paidopen sourcecliapilinuxmac
66new
evidence / 100
2
Docling Open source
Turns PDFs, Office files and scans into structured markdown or JSON, layout and tables intact.

+ Reads a table as a table, and does it on your own machine with nothing sent anywhere.
− Heavy scanned documents want a GPU; CPU-only runs are slow on long PDFs.

MIT; runs locally, no API keyopen sourcecliapilinuxmacwindows
64new
evidence / 100
3
LlamaParse Free tier
Parsing service aimed at the documents that break parsers: nested tables, forms, figures.

+ Holds up on financial statements and forms where plain text extraction turns to soup.
− Per-page pricing and a cloud round trip; large archives add up fast.

Free pages each day, then per-page creditsapiweb
18new
evidence / 100
4
Mathpix Free tier
Screenshots and PDFs of maths and chemistry into LaTeX, markdown or Word.

+ Still the most accurate thing there is at equations, which general OCR gets wrong.
− Narrow by design - a plain business PDF is better served elsewhere.

Free snips each month; paid by volume or API callwebapimacwindowslinuxios
11new
evidence / 100
5
Nanonets Paid
Extracts fields from invoices, receipts and forms, then posts them into an accounting system.

+ Ships the boring half: approval steps, validation rules and exports to QuickBooks or SAP.
− Priced for finance teams, not for a side project.

Per-page tiers; a free trial allowancewebapi
2new
evidence / 100

Work on Surya?

Put your rank on your own site

Free, no attribution deal, no strings. It updates itself every night, and if you drop a place it says so - that is the point of it being worth anything.

Ranked #3 of 7 for extract data from documents on AI Use Cases

HTML

<a href="https://aiusecases.dev/tool/surya/"><img src="https://aiusecases.dev/badge/surya.svg" alt="Ranked #3 of 7 for extract data from documents on AI Use Cases" width="268" height="60"></a>

Markdown

[![Ranked #3 of 7 for extract data from documents on AI Use Cases](https://aiusecases.dev/badge/surya.svg)](https://aiusecases.dev/tool/surya/)

Or buy the box above the list

This page is generated from public signals and nobody can edit it - but you can put a labelled box at the top of Extract data from documents, where people are choosing. 29 USDC for 7 days, one slot, no account, and it never moves a single rank. Sponsor that page, or see how the score is built if a number here looks wrong.