Surya
Surya is a 650M-parameter OCR model for recognizing text, analyzing page layout and reading order, and identifying tables.
What Surya is
Surya is a 650M-parameter OCR model for recognizing text, analyzing page layout and reading order, and identifying tables. The project describes support for more than 90 languages and focuses on documents rather than photos or natural scenes. Surya 2 returns page content in structured blocks that can include HTML, polygons, bounding boxes, confidence values, and reading order. Table recognition can provide row and column geometry; full mode can return HTML with spanning cells and header rows. Equations in page OCR are represented as KaTeX-compatible LaTeX inside math tags. The project offers a Python package, command-line tools, and an interactive Streamlit app for images and PDFs. Layout, OCR, and table recognition require an inference backend: vLLM on NVIDIA GPUs or llama.cpp on CPU and Apple Silicon, or a connection to an existing server. The code is Apache 2.0, but the model weights use a separate modified AI Pubs Open Rail-M license with commercial-use conditions described in the README. Datalab also offers a managed platform, a free public playground, and $5 in free credits.
Good
- Supports OCR in more than 90 languages.
- Returns structured page content and reading order.
- Table output can include row and column geometry.
- Offers Python, command-line, and Streamlit interfaces.
- Handwriting OCR is listed.
Watch out
- Not specialized for photos or natural scenes.
- Layout, OCR, and table recognition need an inference backend.
- Model weights have separate commercial-use conditions.
Who it suits
Surya suits developers and teams building document OCR workflows who can provide a supported inference backend. It is less suited to photo or natural-scene recognition, which is outside its stated specialization.
Verdict
Surya offers detailed document OCR output, including layout, tables, and equations, through several interfaces. Check backend requirements and the separate model-weight license conditions before choosing it for a project.
Plans and prices
From Surya's own pricing page, checked 9 Oct 2026.
| Plan | Price | What it gives you |
|---|---|---|
| Free | Free | $20 monthly allowance on work email or $10 on personal email · 25 requests/min · 25 concurrency · community support · $0 pay-as-you-go |
| Team | $400 a month | $400 monthly usage included, then pay-as-you-go · 400 requests/min · 400 concurrency · email + Slack support · $400 per month |
| Enterprise | Priced on request | Custom usage, rate limits, and terms · dedicated support · VPC or fully air-gapped deployment · Custom |
The facts
Each one read from Surya's own pages, with the date it was checked.
The company
- Headquarters
- Manhattan, New York, United Stateschecked 28 Sep 2026
What it does
- Inference backends
- Layout, OCR, and table recognition use vLLM on NVIDIA GPUs or llama.cpp on CPU and Apple Silicon, and can connect to an existing server.checked 9 Oct 2026
- Installation
- The README installs Surya with pip install surya-ocr and lists Docker plus NVIDIA Container Toolkit for NVIDIA GPU use, or llama.cpp for CPU and Apple Silicon.checked 9 Oct 2026
- Intended users
- Datalab says its platform serves AI/ML teams, enterprises, and product teams building AI agents, RAG systems, and document automation workflows.checked 9 Oct 2026
- Interfaces
- The project provides a Python package, command-line tools, and an interactive Streamlit app for images and PDFs.checked 9 Oct 2026
- Languages
- The repository describes Surya as supporting OCR in 90+ languages.checked 9 Oct 2026
- License
- The Surya code is Apache 2.0, while the model weights use a modified AI Pubs Open Rail-M license with commercial-use conditions described in the README.checked 9 Oct 2026
- Limitations
- Surya is specialized for document OCR rather than photos or natural scenes, and layout, OCR, and table recognition require a running inference backend.checked 9 Oct 2026
- Managed access
- Datalab offers a managed platform running Surya and Chandra, with $5 in free credits and a free public playground.checked 9 Oct 2026
- Math
- Surya 2 recognizes equations inline in page OCR and returns them in KaTeX-compatible LaTeX inside math tags.checked 9 Oct 2026
- Output
- Surya 2 returns page content as structured blocks with HTML, polygons, bounding boxes, confidence, and reading order.checked 9 Oct 2026
- Purpose
- Surya is a 650M-parameter OCR model for text recognition, layout analysis, reading order, and table recognition.checked 9 Oct 2026
- Security and compliance
- Datalab's pricing page lists a SOC 2 Type II report for every plan; BAA/DPA support is listed for Team and Enterprise.checked 9 Oct 2026
- Support
- Datalab lists community support on Free, email and Slack support on Team, and dedicated support on Enterprise.checked 9 Oct 2026
- Supported inputs
- bothchecked 23 Sep 2026
- Supported languages
- 91checked 23 Sep 2026
- Table recognition
- Table recognition can return row and column geometry and, in full mode, HTML with spanning cells and header rows.checked 9 Oct 2026
Where it runs
- Primary platform
- desktopchecked 23 Sep 2026
Alternatives to Surya
- 1
FreeOCR.AI
Free planWebAPImacOSWindows8.2From $80 a year - 2
Tesseract OCR
Free planAndroidAPILinuxmacOS7.9Free plan - 3
Open-Capture
Free planOpen sourceWebAPILinuxSelf-hosted7.9Free plan - 4
ScanScan
AndroidiOSWebmacOS7.7CN¥30 one-time - 5
Prizmo
Free planiOSmacOS7.7Free plan - 6
Tungsten OmniPage
APILinuxmacOSSelf-hosted7.6€99 one-time