Chandra vs Unstructured
Pick Chandra when the job is reading scanned or multilingual pages and tables into Markdown, HTML or JSON for RAG. Pick Unstructured when you ingest many file types (PDF, Word, PowerPoint, email, HTML, images) and need one pipeline with connectors. The main split is that Chandra is a vision-language OCR model, while Unstructured is an ETL library that sorts files into typed elements and gets its best table accuracy only on the hosted platform.
Chandradatalab/chandra | Unstructuredunstructured/unstructured | |
|---|---|---|
| Type | Open source | OCR API |
| Price | 2 credits / page | 20 credits / page |
| Per 1,000 pages | $2.00 | $20.00 |
| Typical speed | 2.0 s / page | 3.0 s / page |
| Best for | Complex layouts and handwriting | Mixed file types into RAG |
| Inputs | PDF, images | PDF, images, DOCX |
| Languages | 40+ languages | Multilingual |
| Max pages per request | 500 | 1,000 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | No |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | No | No |
| Forms | ✓ Yes | No |
| Multilingual | ✓ Yes | ✓ Yes |
| On OCRRouter | Available | Available |
Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.
Which to choose
Choose Chandra if
- Its authors report 85.8 on olmOCR-bench and a 77.8% average across 43 languages, so it suits multilingual scans.
- It outputs Markdown, HTML and JSON with bounding boxes and labels per block, which fits PDF to Markdown for RAG.
- The weights are open and run on a single NVIDIA GPU via vLLM, and the hosted Datalab API costs $1, $3 or $7.50 per 1,000 pages.
Choose Unstructured if
- It handles PDFs, HTML, Word, PowerPoint, emails and images in one library, which suits mixed file ingestion.
- The library is Apache-2.0 with many connectors, and the hosted platform gives the first 10,000 pages free.
- Its Fast strategy suits born-digital PDFs, and Auto, Fast, High-Res and VLM strategies let you trade speed for quality.
Key differences
| Chandra | Unstructured | |
|---|---|---|
| What it is | Document OCR vision-language model from Datalab | Document ETL library and platform that partitions files into typed elements |
| Licence | Code Apache-2.0; weights modified OpenRAIL-M, free for research, personal use and companies under $2M funding or revenue | Apache-2.0 library; proprietary hosted platform |
| Price | Self-hosted free within licence terms; Datalab API $1, $3 or $7.50 per 1,000 pages; $20/month free tier | Library free; hosted first 10,000 pages free, then $0.015 per page (about $15 per 1,000) |
| Output formats | Markdown, HTML, JSON, bounding boxes | JSON, HTML, plain text, bounding boxes |
| Tables | Documents with tables, forms and math are a stated fit | Vendor benchmark of table cell content: 0.426 open-source vs 0.866 hosted; hi_res tables can cut off edges and drop first letters |
| Languages | 77.8% average across 43 languages; 43.2% on Urdu and 51.1% on Sanskrit | Table OCR ignores language settings, so non-English tables are a weak spot |
| Known failure | Token-loop hallucinations reported on 5-6 of about 60 Russian and Kazakh scans; Old Scans category scores 51.1 on olmOCR-bench | Extra Tesseract pass on tables adds processing time; library sends usage telemetry by default |
Sources
- Chandra GitHub README
- Datalab pricing
- GitHub issue #51 token loops
- Unstructured pricing
- Unstructured GitHub README
- Issue #1694
What each one is
Chandra
Datalab's OCR model that keeps layout and handles forms, tables and handwriting.
Chandra detailsUnstructured
Document ETL for RAG. Partitions many file types into typed elements; open-source library too.
Unstructured detailsTry both on your own file
Run the file once with each model and compare the output.
Use one, fall back to the other
{"model": "datalab/chandra","fallbacks": ["unstructured/unstructured"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Chandra alternativesUnstructured alternativesBest OCR models in 2026