- Home
- OCR software
- Surya
Surya
Manhattan, New York, United States
What it is
Surya is a free, open-source 650M-parameter model for document OCR, layout analysis, reading order, and table recognition. It supports OCR in more than 90 languages and is intended for documents rather than photos or natural scenes. Surya 2 returns page content as structured blocks that can include HTML, polygons, bounding boxes, confidence values, and reading order. Table recognition can provide row and column geometry, with full mode returning HTML for spanning cells and header rows. Equations are represented as KaTeX-compatible LaTeX inside math tags. The project offers a Python package, command-line tools, and a Streamlit interface for images and PDFs. Layout, OCR, and table features require an inference backend: vLLM on NVIDIA GPUs or llama.cpp on CPU and Apple Silicon, or an existing server. The code uses Apache 2.0, while model weights have separate commercial-use conditions. Datalab also offers managed access, a public playground, and $5 in free credits.
For and against
For
- Supports more than 90 OCR languages
- Structured output includes reading order
- Table geometry and HTML output
- Python, CLI, and Streamlit interfaces
- Free and open source
Against
- Focused on documents, not natural scenes
- Inference backend must be running
- Model weights have separate license conditions
The verdict
Surya provides broad document OCR and structured layout output through several developer interfaces. Users should plan for backend setup and review the model-weight licensing terms before commercial deployment.
Who it suits
Surya suits developers building document-processing, RAG, or automation workflows who can operate the required inference backend. It supports Linux, macOS, API use, and self-hosting.
Plans and prices
Free
$20 monthly allowance on work email or $10 on personal email · 25 requests/min · 25 concurrency · community support
Read 9 Oct 2026 · datalab.to
Team
$400 monthly usage included, then pay-as-you-go · 400 requests/min · 400 concurrency · email + Slack support
Read 9 Oct 2026 · datalab.to
Enterprise
Custom usage, rate limits, and terms · dedicated support · VPC or fully air-gapped deployment
Read 9 Oct 2026 · datalab.to
On the record
- 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.github.com · read 9 Oct 2026
- Supported inputs
- bothgithub.com · read 23 Sep 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.github.com · read 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.documentation.datalab.to · read 9 Oct 2026
- Interfaces
- The project provides a Python package, command-line tools, and an interactive Streamlit app for images and PDFs.github.com · read 9 Oct 2026
- Supported languages
- 91github.com · read 23 Sep 2026
- Languages
- The repository describes Surya as supporting OCR in 90+ languages.github.com · read 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.github.com · read 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.github.com · read 9 Oct 2026
- Managed access
- Datalab offers a managed platform running Surya and Chandra, with $5 in free credits and a free public playground.github.com · read 9 Oct 2026
- Math
- Surya 2 recognizes equations inline in page OCR and returns them in KaTeX-compatible LaTeX inside math tags.github.com · read 9 Oct 2026
- Output
- Surya 2 returns page content as structured blocks with HTML, polygons, bounding boxes, confidence, and reading order.github.com · read 9 Oct 2026
- Primary platform
- desktopgithub.com · read 23 Sep 2026
- Purpose
- Surya is a 650M-parameter OCR model for text recognition, layout analysis, reading order, and table recognition.github.com · read 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.datalab.to · read 9 Oct 2026
- Support
- Datalab lists community support on Free, email and Slack support on Team, and dedicated support on Enterprise.datalab.to · read 9 Oct 2026
- Table recognition
- Table recognition can return row and column geometry and, in full mode, HTML with spanning cells and header rows.github.com · read 9 Oct 2026
