Datalab

State of the Art models for Document Intelligence

---

Surya

Surya is a document OCR toolkit built around a single vision-language model that does:

• Full-page OCR with layout, ranking near top models on olmOCR-bench
• Line-level text detection
• Layout analysis (table, image, header, etc.) with reading order
• Table recognition (rows + columns + cell HTML)
• Math / equations recognized inline (no separate LaTeX OCR pass)

It works on a range of documents (see usage and benchmarks).

Try Datalab's Managed Platform

Our managed platform runs both Surya, and variants of our highest accuracy model, Chandra.

If you have high volume workloads, we offer a batch processing service that can process 1B+ pages per week.

Get started with $5 in free credits — sign up (takes under 30 seconds) or try our public playground.

Commercial self-hosting of the model weights requires a license — see Commercial usage. For on-prem licensing, contact us.

Detection	OCR

Layout	Reading Order

Table Recognition	Math / Equations

Surya is named for the Hindu sun god, who has universal vision.

Examples

Each row links to four annotated views of the same page: text-line detection, layout, reading order, and (when present) table recognition.

Name	Detection	Layout	Order	Table Rec
Newspaper	Image	Image	Image
Textbook	Image	Image	Image
Tax Form	Image	Image	Image	Image
Handwritten Notes	Image	Image	Image	Image
Corporate Doc	Image	Image	Image	Image

Commercial usage

The Surya code is licensed under Apache 2.0. The model weights use a modified AI Pubs Open Rail-M license (free for research, personal use, and startups under $2M funding/revenue). For broader commercial licensing of the model weights, visit our pricing page here.

Installation

You'll need python 3.10+ and PyTorch. You may need to install the CPU version of torch first if you're not using a Mac or a GPU machine. See here for more details.

Install with:

pip install surya-ocr

Upgrading from Surya v1

Surya 2 replaces the per-task encoder-decoder models (FoundationPredictor + RecognitionPredictor + LayoutPredictor + TableRecPredictor each holding their own torch checkpoints) with a single vision-language model served by vllm (Docker, GPU) or llama-server (Apple Silicon / CPU). If you have v1 code, you can migrate to this:

# v2
from surya.inference import SuryaInferenceManager
from surya.recognition import RecognitionPredictor

manager = SuryaInferenceManager()              # auto-spawns vllm or llama-server
rec = RecognitionPredictor(manager)
predictions = rec([image])                     # full-page OCR by default

What's different:

• SuryaInferenceManager replaces FoundationPredictor. Same manager instance is shared across LayoutPredictor, RecognitionPredictor, TableRecPredictor.
• Output schemas changed: see the per-section JSON tables below. Highlights — text_lines → blocks (with html); layout dropped top_k, added count; table_rec dropped is_header / colspan / rowspan from cells.

Usage

Surya 2 runs layout, OCR, and table recognition through a single VLM served by vllm (GPU) or llama.cpp (CPU / Apple Silicon). The inference manager will spawn one for you on first use; you can also point it at an existing server via SURYA_INFERENCE_URL=http://host:port/v1.

• Inspect the settings in surya/settings.py. You can override any setting via env var (e.g. SURYA_INFERENCE_BACKEND=vllm).
• Text detection and OCR errors are separate models.

Interactive App

I've included a streamlit app that lets you interactively try Surya on images or PDF files. Run it with:

pip install streamlit pdftext
surya_gui

OCR (text recognition)

This command will write out a json file with the detected text and bboxes:

surya_ocr DATA_PATH

• DATA_PATH can be an image, pdf, or folder of images/pdfs
• --images will save images of the pages and detected blocks (optional)
• --output_dir specifies the directory to save results to instead of the default
• --page_range specifies the page range to process in the PDF, specified as a single number, a comma separated list, a range, or comma separated ranges - example: 0,5-10,20.

The results.json file contains a dict keyed by input filename (no extension). Each value is a list of page dicts. Each page dict contains:

• blocks - per-block OCR results in reading order
  • label - canonicalized layout label (e.g. Text, Section-Header, Table, Equation-Block, Picture, Form, Page-Header, ...). See surya/inference/prompts.py:LAYOUT_LABEL_SET for the full set.
  • raw_label - original label emitted by the model, before canonicalization
  • reading_order - 0-indexed position in layout output
  • html - block content as HTML (math wrapped in <math>...</math>, tables as <table>...</table>, etc.). "" if the block was skipped
  • polygon - 4-corner polygon in [[x0,y0],[x1,y0],[x1,y1],[x0,y1]] order
  • bbox - axis-aligned [x0, y0, x1, y1] derived from the polygon
  • confidence - mean per-token probability across the block's decode (0-1)
  • skipped - true if the block was a visual label (e.g. Picture) and not OCR'd
  • error - true if the block OCR call failed
• image_bbox - [0, 0, width, height] for the page image

Performance tips

Throughput is governed by the inference backend, not a RECOGNITION_BATCH_SIZE env var. With vllm, raise --max-num-seqs / --max-num-batched-tokens (or SURYA_INFERENCE_PARALLEL on the client side) to keep more pages in flight. With llama.cpp, set SURYA_INFERENCE_PARALLEL to match --parallel on llama-server.

From python

from PIL import Image
from surya.inference import SuryaInferenceManager
from surya.recognition import RecognitionPredictor

manager = SuryaInferenceManager()
recognition_predictor = RecognitionPredictor(manager)

# Default: full-page OCR. One VLM call per page; returns layout + content as
# HTML <div data-bbox=... data-label=...> blocks.
predictions = recognition_predictor([Image.open(IMAGE_PATH)])

# Block mode: pre-run layout, then per-block OCR. Auto-selected when
# `layout_results` is passed.
from surya.layout import LayoutPredictor
layout = LayoutPredictor(manager)
layouts = layout([Image.open(IMAGE_PATH)])
predictions = recognition_predictor([Image.open(IMAGE_PATH)], layouts)

Text line detection

This command will write out a json file with the detected bboxes.

surya_detect DATA_PATH

• DATA_PATH can be an image, pdf, or folder of images/pdfs
• --images will save images of the pages and detected text lines (optional)
• --output_dir specifies the directory to save results to instead of the default
• --page_range specifies the page range to process in the PDF, specified as a single number, a comma separated list, a range, or comma separated ranges - example: 0,5-10,20.

The results.json file will contain a json dictionary where the keys are the input filenames without extensions. Each value will be a list of dictionaries, one per page of the input document. Each page dictionary contains:

• bboxes - detected bounding boxes for text
  • bbox - the axis-aligned rectangle for the text line in (x1, y1, x2, y2) format. (x1, y1) is the top left corner, and (x2, y2) is the bottom right corner.
  • polygon - the polygon for the text line in (x1, y1), (x2, y2), (x3, y3), (x4, y4) format. The points are in clockwise order from the top left.
  • confidence - the confidence of the model in the detected text (0-1)
• vertical_lines - vertical lines detected in the document
  • bbox - the axis-aligned line coordinates.
• page - the page number in the file
• image_bbox - the bbox for the image in (x1, y1, x2, y2) format. (x1, y1) is the top left corner, and (x2, y2) is the bottom right corner. All line bboxes will be contained within this bbox.

Performance tips

Detection is a torch model. DETECTOR_BATCH_SIZE (default 36) controls VRAM usage on GPU; raise it on larger cards.

From python

from PIL import Image
from surya.detection import DetectionPredictor

det_predictor = DetectionPredictor()
predictions = det_predictor([Image.open(IMAGE_PATH)])

Layout and reading order

This command will write out a json file with the detected layout and reading order.

surya_layout DATA_PATH

• DATA_PATH can be an image, pdf, or folder of images/pdfs
• --images will save images of the pages and detected text lines (optional)
• --output_dir specifies the directory to save results to instead of the default
• --page_range specifies the page range to process in the PDF, specified as a single number, a comma separated list, a range, or comma separated ranges - example: 0,5-10,20.

The results.json file contains a dict keyed by input filename (no extension). Each value is a list of page dicts. Each page dict contains:

• bboxes - layout boxes in reading order
  • polygon - 4-corner polygon [[x0,y0],[x1,y0],[x1,y1],[x0,y1]]
  • bbox - axis-aligned [x0, y0, x1, y1] derived from the polygon
  • label - canonicalized label. One of Caption, Footnote, Equation-Block, List-Group, Page-Header, Page-Footer, Image, Section-Header, Table, Text, Complex-Block, Code-Block, Form, Table-Of-Contents, Figure, Chemical-Block, Diagram, Bibliography, Blank-Page
  • raw_label - original label emitted by the model
  • position - 0-indexed reading order
  • count - model's token estimate for OCR'ing this block (rounded to multiples of 50; used to size the per-block decode budget)
  • confidence - mean per-token probability across the layout decode (0-1)
• image_bbox - [0, 0, width, height]
• raw - raw JSON the layout model emitted, for debugging
• error - true if the layout call failed

Performance tips

Layout runs through the shared inference backend. Throughput tuning is the same as OCR — see Performance tips above.

From python

from PIL import Image
from surya.inference import SuryaInferenceManager
from surya.layout import LayoutPredictor

layout_predictor = LayoutPredictor(SuryaInferenceManager())
layout_predictions = layout_predictor([Image.open(IMAGE_PATH)])

Table Recognition

This command will write out a json file with the detected table cells and row/column ids, along with row/column bounding boxes. If you want to get cell positions and text, along with nice formatting, check out the marker repo. You can use the TableConverter to detect and extract tables in images and PDFs. It supports output in json (with bboxes), markdown, and html.

surya_table DATA_PATH

• DATA_PATH can be an image, pdf, or folder of images/pdfs
• --images will save annotated row + column overlays alongside the json (optional)
• --output_dir specifies the directory to save results to instead of the default
• --page_range specifies the page range to process in the PDF, specified as a single number, a comma separated list, a range, or comma separated ranges - example: 0,5-10,20.
• --skip_table_detection tells table recognition not to detect tables first. Use this if your image is already cropped to a table.

The results.json file contains a dict keyed by input filename (no extension). Each value is a list of per-table dicts. Each table dict contains:

• rows - detected table rows in reading order
  • polygon / bbox - row geometry (same convention as everywhere else)
  • row_id - 0-indexed row id
• cols - detected table columns
  • polygon / bbox - column geometry
  • col_id - 0-indexed column id
• cells - geometric row × column intersections (simple mode)
  • polygon / bbox - cell geometry
  • row_id, col_id, cell_id
• html - full <table>...</table> HTML (only populated when predict_full is used; handles spanning cells / header rows). null in simple mode.
• mode - "simple" or "full"
• image_bbox - the table crop bbox
• error - true if the table_rec call failed
• raw - raw model output, for debugging

Performance tips

Table recognition routes through the shared VLM. Throughput tuning is the same as OCR.

From python

from PIL import Image
from surya.inference import SuryaInferenceManager
from surya.table_rec import TableRecPredictor

table_rec_predictor = TableRecPredictor(SuryaInferenceManager())

# Default: rows + columns only, cells derived from intersections.
table_predictions = table_rec_predictor([Image.open(IMAGE_PATH)])

# Or full HTML output (better for spanning cells / headers):
# table_predictions = table_rec_predictor.predict_full([image])

Math / equations

Surya 2 handles math inline as part of full-page OCR — recognized equations come back inside <math>...</math> tags in the same HTML output as surrounding prose, in KaTeX-compatible LaTeX. No separate LaTeX OCR pass.

Inference Backends

Layout / OCR / table_rec all share one VLM, served either by vllm (GPU) or llama.cpp (CPU / Apple Silicon). The SuryaInferenceManager will spawn one automatically; you can also point at a pre-running server:

# Attach to an existing vllm
export SURYA_INFERENCE_BACKEND=vllm
export SURYA_INFERENCE_URL=http://localhost:8000/v1

Setting	Default	Notes
SURYA_INFERENCE_BACKEND	auto (vllm if NVIDIA, else llamacpp)	vllm | llamacpp | unset (auto)
SURYA_INFERENCE_URL	(auto-spawn)	Attach to a running OpenAI-compatible server
SURYA_INFERENCE_PARALLEL	8	Client-side concurrency to the backend
SURYA_GUIDED_LAYOUT	true	JSON-schema-constrained layout decode

Limitations

• This is specialized for document OCR. Performance on photos or natural scenes is not the goal.
• Layout / OCR / table_rec all need a running inference backend (vllm or llama.cpp). Detection runs purely on torch and works without it.

Troubleshooting

If OCR isn't working properly:

• Try increasing resolution of the image so the text is bigger. If the resolution is already very high, try decreasing it to no more than a 2048px width.
• Preprocessing the image (binarizing, deskewing, etc) can help with very old/blurry images.
• You can adjust DETECTOR_BLANK_THRESHOLD and DETECTOR_TEXT_THRESHOLD if you don't get good results. DETECTOR_BLANK_THRESHOLD controls the space between lines - any prediction below this number will be considered blank space. DETECTOR_TEXT_THRESHOLD controls how text is joined - any number above this is considered text. DETECTOR_TEXT_THRESHOLD should always be higher than DETECTOR_BLANK_THRESHOLD, and both should be in the 0-1 range. Looking at the heatmap from the debug output of the detector can tell you how to adjust these (if you see faint things that look like boxes, lower the thresholds, and if you see bboxes being joined together, raise the thresholds).

Manual install

If you want to develop surya, you can install it manually with uv:

git clone https://github.com/VikParuchuri/surya.git
cd surya
uv sync --group dev      # installs runtime + dev deps
uv run surya_ocr ...     # or `uv shell` to enter the venv

Benchmarks

Surya 2 is a single VLM that handles layout analysis, OCR (full-page or per-block), and table recognition in one model. We evaluate end-to-end on olmOCR-bench — the standard quality benchmark for document parsers.

olmOCR-bench

Model	Params	Score
Infinity-Parser2-Pro	35.1B	87.6
Chandra OCR 2 (Datalab)	5.3B	85.9
dots.mocr	3.0B	83.9
LightOnOCR 2-1B *	1.0B	83.2
Surya OCR 2 (Datalab)	0.69B	83.1
Chandra OCR 1 (Datalab)	9.0B	83.1
olmOCR (anchored)	8.3B	77.4
GOT OCR	0.6B	48.3

* LightOnOCR 2-1B uses a different benchmark methodology than the other entries (see their release notes); the score is included for context but is not directly comparable.

Comparison scores from the olmOCR-bench dataset card. Surya OCR 2 is reported as 0.69B params — the on-disk safetensors duplicates the tied embedding + lm_head, so HuggingFace shows ~0.75B; the underlying parameter count is 0.69B.

Surya 2, per-source pass rate on the default preset (8,413 tests total):

ArXiv	Base	Hdr/Ftr	TinyTxt	MultCol	OldScan	OldMath	Tables
88.7	99.9	92.1	86.4	82.6	42.8	85.8	86.6

Throughput

Full-page OCR, 96 DPI input (~3,000 output tokens/page average), measured client-side against a running inference server.

B200 (vllm)

vllm/vllm-openai:v0.20.1, single B200, --max-num-seqs 512 / --max-num-batched-tokens 16384. Each row uses ≥20× concurrency in pages so steady state dominates. Prefix caching off.

Concurrency	Pages	Wall (s)	Pages/s	Tokens/s	p50 (ms)	p95 (ms)	p99 (ms)
64	1,280	270	4.74	14,058	13,806	18,656	19,217
128	2,560	449	5.71	16,948	22,820	29,501	30,897
256	5,120	882	5.81	17,186	44,253	51,119	52,333
512	10,240	1,761	5.81	17,211	87,961	94,941	96,185

Throughput saturates at conc=256 (5.81 pages/s ≈ 17.2k tokens/s). Going to 512 doesn't add capacity — it just queues, doubling latency. For production batch jobs the sweet spot is conc=128–256.

Apple Silicon (llama.cpp / Metal)

llama-server with Metal backend, one process per --parallel level.

--parallel	Pages/s	Tokens/s	p50 (ms)	p95 (ms)
4	0.217	222	18,122	19,676
8	0.269	270	27,549	42,730
16	0.264	286	53,166	70,378

Knees at --parallel=8 — Metal is decode-saturated past that point.

Reproducing

We score Surya 2 on olmOCR-bench by serving the model with vllm (or llama.cpp) and running the official olmOCR-bench harness from allenai/olmocr. Use the HIGH_ACCURACY_BBOX_PROMPT prompt (single full-page call per page); the RecognitionPredictor defaults to that mode.

Training

Layout, OCR, and table recognition all share a single vision-language model (Qwen3.5-style architecture, ~770M params). It's trained on diverse document images to emit either a layout JSON or a full-page HTML output, depending on prompt. Text-line detection is a separate small torch model — a modified EfficientViT segformer trained from scratch on document line annotations.

If you want help finetuning Surya on your own data, or to use our managed training stack, reach us at hi@datalab.to.

Thanks

This work would not have been possible without amazing open source AI work:

• Qwen3-VL from Alibaba (architecture basis for the Surya 2 VLM)
• vllm and llama.cpp for inference
• Segformer from NVIDIA
• EfficientViT from MIT
• timm from Ross Wightman
• transformers from huggingface
• CRAFT, a great scene text detection model

Thank you to everyone who makes open source AI possible.

Citation

If you use surya (or the associated models) in your work or research, please consider citing us using the following BibTeX entry:

@misc{paruchuri2025surya,
  author       = {Vikas Paruchuri and Datalab Team},
  title        = {Surya: A lightweight document OCR and analysis toolkit},
  year         = {2025},
  howpublished = {\url{https://github.com/VikParuchuri/surya}},
  note         = {GitHub repository},
}