LlamaParse vs docTR
Pick LlamaParse when you need PDFs and office files turned into Markdown or JSON for RAG and can accept a hosted, proprietary API. Pick docTR when you need word-level OCR with coordinates that you run yourself under Apache-2.0, for example on invoices and receipts. The main split is output: LlamaParse gives structured documents, docTR gives only words and lines.
LlamaParsellamaindex/llamaparse | docTRmindee/doctr | |
|---|---|---|
| Type | OCR API | Open source |
| Price | 5 credits / page | 1 credit / page |
| Per 1,000 pages | $5.00 | $1.00 |
| Typical speed | 3.0 s / page | 0.9 s / page |
| Best for | RAG ingestion | Word-level OCR with positions |
| Inputs | PDF, images, DOCX | PDF, images |
| Languages | 80+ languages | Mainly Latin scripts |
| Max pages per request | 1,000 | 500 |
| Tables | ✓ Yes | No |
| Handwriting | No | No |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | No | ✓ Yes |
| Forms | No | No |
| Multilingual | ✓ Yes | No |
| 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 LlamaParse if
- You want Markdown or JSON from mixed PDF and office documents for RAG, including inside LlamaIndex pipelines.
- You want to trade cost for quality per job, from 1 credit per page in Fast mode up to 45 credits in Agentic Plus.
- You want a free start: 10,000 credits per month, and re-parsing the same file within 48 hours costs nothing because results are cached.
Choose docTR if
- You need word-level text with bounding boxes, in a nested page, block, line and word JSON structure.
- You want to run OCR on your own CPU or NVIDIA GPU, with Apache-2.0 licensing, Docker images and no per-page fee.
- Your team writes Python and wants to choose detection and recognition architectures, or train its own models, with a key information extraction predictor for invoices and receipts.
Key differences
| LlamaParse | docTR | |
|---|---|---|
| How you run it | Hosted API from LlamaIndex (LlamaCloud) | Python library on PyTorch, run locally or in Docker |
| Licence | Proprietary | Apache-2.0 |
| Price | $1.25 per 1,000 credits; about $1.25 to $56 per 1,000 pages depending on mode; 10K free credits per month | Free and self-hosted; you pay for compute |
| Output formats | Markdown, JSON, plain text, images | JSON, plain text, bounding boxes |
| Tables and layout | Tiers up to LLM-based agentic parsing for hard layouts | No table, formula or Markdown output; returns words and lines only |
| Hardware | None to manage; runs in the hosted service | CPU or NVIDIA GPU; GPU Docker images need CUDA 12.2 or newer on the host |
| Known failure | Hallucinates on low-quality scans with no confidence score (issue #483); invented transactions on bank statements in Accurate mode (issue #450); a job stuck in progress for days (issue #623) | Split-crop merging misjoins readings across repeated characters (issue #2229); cropping images gives poorer results (issue #2007); ONNX export of MASTER truncates decoding steps (issue #2224) |
Sources
- LlamaIndex pricing
- GitHub issue #450
- GitHub issue #483
- docTR GitHub README
- Issue #2229
- docTR documentation
What each one is
docTR
Open-source OCR library from Mindee with separate text detection and recognition models.
docTR detailsTry 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": "llamaindex/llamaparse","fallbacks": ["mindee/doctr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
LlamaParse vs KrakenLlamaParse vs RapidOCRMathpix Convert vs docTRClaude vs docTRLlamaParse vs Reducto ParseLlamaParse vs DoclingdocTR vs OCR.spacedocTR vs Typhoon OCR
LlamaParse alternativesdocTR alternativesBest OCR models in 2026