Nanonets

Your LLMs are hungry for data, but documents are messy 😩. DocStrange is the answer! Our open-source solution turns any document into clean, LLM-ready data with one command. Give your models what they need. 🔗 github.com/NanoNets/docst… 🔗 pypi.org/project/docstr…

More info here -> nanonets.com/research/nanon… Try API -> docstrange.nanonets.com/app

Nanonets OCR-3 is live. This is the most accurate OCR model in the world currently. 87.4 on OLM-OCR (Global #1) 85.9 on IDP Leaderboard (Global #1) 90.5 on OmniDocBench OCR-3 also ships with two critical features that foundational models and VLMs miss today - confidence scores and bounding boxes.

Nanonets OCR-3 is the only OCR model you'll need in your agentic stack. The model API exposes five endpoints - /parse - structured markdown /extract - structured outputs in your schema /split - classify or split outputs based on content /chunk - context-aware chunks optimized for RAG /vqa - grounded answers with bboxes over sources We've specifically fine-tuned the model on edge cases where OCR repeatedly fails - complex tables, forms, non-trivial layouts.

With bounding boxes, you get exact coordinates for every extracted element. Use them for - 1. RAG citations 2. Feeding specific document regions to agents 3. Agent observability With confidence scores, you can measure reliability of every extraction. Pass high-confidence outputs directly, route low-confidence outputs to human review or a larger model. Use them to push your net accuracy to near 100%.

#changelog Introducing AI Agent Guidelines. 👉 changelog.nanonets.com/introducing-ai…

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

nanonets-ocr-s を vision-language.cpp でスマッホで動かしたいから優秀な VLM 若人さまはよじゃぶじゃぶ湧き出てきてもろて？🥺👊

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

@mervenoyann AWS Textract and Nanonets were on par with each other.

a question for y'all: which PDF renderers do you use (other than Docling, SmolDocling and R/OlmOCR)? why do you prefer that over these ones? 👀

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

I just published Nanonets OCR: A Small Gem for Handwritten Notes. As always, the testing was done with mlx-vlm. medium.com/p/nanonets-ocr… #mlx #mlx_vlm #OCR

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

Does anyone have any contacts in Nanonets - Bangalore ??

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

Just tried the new Nanonets-OCR-s. Works really well! - Small size: just 3.75B parameters, works on RTX 3060 without quantization. - Recognizes equations and tables! And tables with equations (!) - Is multilingual! - Outputs descriptions of the images - Outputs in Markdown format

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

Nanonets-OCR-s predstavlja veliki iskorak u odnosu na klasične OCR (optical character recognition – optičko prepoznavanje karaktera) alate. Dok većina OCR sistema samo prepoznaje i transkribuje tekst iz slika, Nanonets-OCR-s strukturira dokumente na način koji je optimizovan za dalju obradu pomoću velikih jezičkih modela (LLM) – kao što su GPT, Claude i Gemini.

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

@alexmcaulay We'd like to throw our hat in the ring

We are testing document parsing engines right now for a major project and going to report back on our findings. We are testing Docling, N8N, MarkITDown, LlamaParse, Mistral, Rossum, Veryfi, Google Document AI, Amazon Textract. Going to give a really good breakdown of everything for you. Anything else we should test?

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

🥳OCRFlux shines in accuracy! It achieves significantly higher Edit Distance Similarity (EDS) and Tree Edit Distance-based Similarity (TEDS) scores compared to #olmOCR and #Nanonets. Try our demo: 2ly.link/28rkt #OCRFlux #OCR #Benchmark

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

Nanonets-OCR-s – OCR model transforms documents into structured markdown huggingface.co/nanonets/Nanon… (news.ycombinator.com/item?id=442870…)

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

@karthikreddy95 @venky4a @miryalasrikanth Nanonets is the best IDP out there in terms of accuracy and pricing. PS: I've personally evaluated their APIs using Indian Regional Invoices (Most toughest out of all)

డాక్టరు అవటం వలన మరో అడ్వాంటేజీ, వేరే డాక్టరు గీకిపడేసిన మాత్రల్ని మనం ఇంట్లోవాళ్లకి అర్థమయ్యేట్లుగా రాసి ఇవ్వటం.

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

📄 Nanonets vient de sortir Nanonets-OCR-s, un modèle IA révolutionnaire qui transforme tes docs (images, PDFs) en Markdown bien structuré. Ça gère les équations LaTeX, tables complexes, signatures et plus encore ! Suivez ce thread pour en savoir plus ! 👇 #NanonetsOCR #AI #Tech

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

@andersonbcdefg @hoanhle_ In my tests I found rolmocr (reducto’s version of olmocr) and nanonets worked best fwiw (much better than marker). Everything whiffs on math expressions despite the latex support

it's shocking how terrible the current state of OCR is given how many companies are working on doc intelligence AND that we are supposedly "almost at AGI"

Introducing Nanonets-OCR2: a lightweight 3B VLM that transforms documents into clean, structured Markdown and is capable of VQA. We have trained the model on close to 3 million documents. It is multi-lingual and can handle handwritten documents. Live demo: docstrange.nanonets.com HF: huggingface.co/nanonets/Nanon… Blog: nanonets.com/research/nanon…

Nanonets does: 🧮 LaTeX Equation Recognition -> Transforms mathematical equations into perfect LaTeX syntax. 🖼️ Intelligent Image Description -> Automatically describes images with structured tags for smooth LLM processing. ✒️ Signature Detection & Isolation -> Accurately identifies and isolates signatures from other text, streamlining legal and business document handling. 💧 Watermark Extraction -> Extracts watermark text seamlessly, keeping content structured and context clear. ☑️ Smart Checkbox Handling -> Converts checkboxes and radio buttons into standardized Unicode symbols (☐, ☑, ☒) for clarity and consistency. 📊 Complex Table Extraction -> Handles intricate tables, converting them into both markdown and HTML.

📢 A new open-source OCR model is breaking the internet: Nanonets-OCR-s! Nanonets understands context and semantic structures, transforming documents into clean, structured markdown. It has an Apache 2.0 license, and the authors compare it to Mistral-OCR 🧵 Let's look closer: