OCRRouter

Claude vs Chunkr

Pick Claude to read, question or extract from long PDFs when you want reasoning and page-level citations. Pick Chunkr when you need bounding boxes, layout-aware chunks for RAG, or a self-hosted parser. The main split is that Claude is a hosted, token-priced language model with no bounding boxes, while Chunkr is a parsing service you can run yourself.

Claudeanthropic/claudeChunkrchunkr/chunkr
TypeVision modelOCR API
Price25 credits / page14 credits / page
Per 1,000 pages$25.00$14.00
Typical speed2.6 s / page3.0 s / page
Best forLong dense documentsLayout-aware chunks for RAG
InputsPDF, imagesPDF, images, DOCX
Languages50+ languagesMultilingual
Max pages per request1001,000
Tables✓ Yes✓ Yes
Handwriting✓ YesNo
Math✓ Yes✓ Yes
Bounding boxesNo✓ Yes
ConfidenceNoNo
FormsNoNo
Multilingual✓ Yes✓ Yes
On OCRRouterAvailableAvailable

Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.

Which to choose

Choose Claude if

  • You want document question answering or summaries of long reports, with page-level citations from the API.
  • You need extraction to JSON with reasoning, and each page is read as both text and a rendered image, so charts and scans are visible.
  • You send large files: up to 600 pages per request on 1M-context models (100 pages otherwise) and a 32 MB payload.

Choose Chunkr if

  • You need OCR with bounding boxes, which Claude is not suited for.
  • You want to self-host with Docker (GPU, CPU-only or Mac ARM) and use any OpenAI-compatible LLM endpoint.
  • You are chunking PDFs, slides, Word files or images for RAG and want HTML, Markdown or JSON output.

Key differences

ClaudeChunkr
How you run itHosted API onlyCloud API or self-hosted open-source version
LicenceProprietaryAGPL-3.0 or commercial licence; closed-source products need the commercial licence
PricingPer token, no extra PDF fee; a PDF page typically uses 1,500 to 3,000 tokens plus image tokensHosted credit-based plans (prices not verified); open-source version is free but you pay for compute and an LLM endpoint
Output formatsMarkdown, JSON, plain textHTML, Markdown, JSON, bounding boxes
Bounding boxesNot suited to pipelines that need themOCR with bounding boxes
Input and limitsPDFs and images; up to 600 pages per request on 1M-context models (100 otherwise), 32 MB; password-protected or encrypted PDFs not supportedPDF, PPT, Word and images; Excel not supported in the open-source version
Known weaknessBulk OCR is costly versus per-page OCR APIs; no published OmniDocBench scoreOpen-source version uses community models that are less accurate than the cloud's; self-host still needs an external LLM endpoint

Sources

What each one is

Claude

General vision model with careful reading of long, dense documents.

Claude details

Chunkr

Document parsing API that segments and chunks pages for RAG. Self-hostable version available.

Chunkr details

Try both on your own file

Run the file once with each model and compare the output.

PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?

Use one, fall back to the other

{"model": "anthropic/claude","fallbacks": ["chunkr/chunkr"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

Claude alternativesChunkr alternativesBest OCR models in 2026