Readable on the first pass
Use markdown when humans still need to scan the output quickly without losing structure like headings, tables, or section breaks.
PDF to markdown API
Convert PDFs and scanned documents into clean markdown with a developer-first OCR API
Use LeapOCR when you need PDF to markdown output that stays readable in real workflows. Keep headings, tables, and sections intact, then add instructions for translation, cleanup, or normalization only when a page needs extra work.
Standard-v1 credit ladder
Readable by default. Controllable when needed.
Base OCR gives you structured markdown with headings, lists, and tables preserved. Custom instructions and bbox stay optional, so you only pay more when the output has to do more.
Standard-v1 pricing
Every step is priced per page. Start with base OCR, then add customization and bbox only when the page calls for it.
Base OCR
1 credit / page
Read the page into clean markdown with headings, tables, and lists kept intact.
Step
01
Customize
+1 credit / page
Apply instructions like translate to French, normalize labels, or collapse noisy sections.
Step
02
Bounding boxes
+1 credit / page
Return coordinates for fields, lines, tables, or signatures when layout matters downstream.
Step
03
What you can ask for
Keep the base flow simple, then add instructions where it matters
Translate section labels to French Normalize totals and dates Condense noisy footer language Attach bbox to key sections
Adapt the output with instructions
Translate, normalize labels, condense noisy sections, or enforce a cleaner structure without changing the ingest pipeline.
Add layout only when it helps
Attach bbox when reviewers, overlays, or internal tools need geometry. Leave it off when plain markdown is enough.
Examples
Five markdown outputs rebuilt from real documents
Each example starts from OCR on a real document image, then shows how markdown can stay useful for review, translation, cleanup, and layout-aware workflows.
Open full image
Invoice
Invoice as markdown with section bbox
Direct OCR from the invoice sample, organized into markdown with semantic section bbox attached to the main document blocks and no instruction layer.
Hard part: Tables plus finance totals
Base OCR Customize BBox
Markdown output with bbox
md
# Invoice INV-100## Seller <!-- bbox section.seller: [0.04, 0.80, 0.24, 0.12] -->- Company: CONTOSO LTD.- Address: Contoso Headquarters, 123 456th St, New York, NY, 10001## Customer <!-- bbox section.customer: [0.57, 0.80, 0.37, 0.11] -->- Customer name: MICROSOFT CORPORATION- Customer ID: CID-12345- Customer address: Microsoft Corp, 123 Other St, Redmond WA, 98052## Billing and delivery <!-- bbox section.billing_and_delivery: [0.05, 0.54, 0.74, 0.13] -->- Bill to: Microsoft Finance, 123 Bill St, Redmond WA, 98052- Ship to: Microsoft Delivery, 123 Ship St, Redmond WA, 98052- Service address: Microsoft Services, 123 Service St, Redmond WA, 98052## Invoice metadata- Invoice date: 11/15/2019- Due date: 12/15/2019- Service period: 10/14/2019 - 11/14/2019- P.O. number: PO-3333## Line items <!-- bbox section.line_items: [0.03, 0.36, 0.91, 0.11] -->| Quantity | Description | Unit price | Total || -------- | ------------------ | ---------- | ------- || 1 | Consulting service | $100.00 | $100.00 |## Totals <!-- bbox section.totals: [0.56, 0.23, 0.39, 0.15] -->- Subtotal: $100.00- Sales tax: $10.00- Total: $110.00- Previous unpaid balance: $500.00- Total due: $610.00## Remit to <!-- bbox section.remit_to: [0.04, 0.09, 0.23, 0.10] -->- Contoso Billing- 123 Remit St- New York, NY, 10001
Open full image
Receipt
Receipt translated to French with section bbox
The receipt is photographed at an angle and includes handwritten values, with semantic bbox returned for the main document sections.
Hard part: Skewed image plus handwritten value
Instruction
Translate section labels to French after OCR, but keep the merchant and item names unchanged.
Base OCR Customize BBox
Markdown output with bbox
md
# Recu <!-- bbox section.header: [0.12, 0.64, 0.62, 0.28] -->- Commercant: Contoso- Adresse: 123 Main Street, Redmond, WA 98052- Telephone: 987-654-3210- Date: 6/10/2019 13:59- Vendeur: Paul## Articles <!-- bbox section.items: [0.11, 0.41, 0.66, 0.19] -->| Qté | Article | Montant || --- | ------------- | ------- || 1 | Cappuccino | $2.20 || 1 | BACON & EGGS | $9.5 || | Sunny-side-up | |## Totaux <!-- bbox section.totals: [0.22, 0.14, 0.58, 0.25] -->- Sous-total: $11.70- Taxe: $1.17- Pourboire: $1.63- Total: $14.50
Open full image
Identity document
Driver license as direct OCR markdown
A compact card layout with dense labels, small text, and several adjacent identity fields.
Hard part: Compact layout and critical identity fields
Base OCR Customize BBox
Markdown output
md
# Driver License- State: WASHINGTON USA- Document type: DRIVER LICENSE- Notice: FEDERAL LIMITS APPLY- License number: WDLABCD456DG- Class: B- Donor: yes- Last name: TALBOT- First name: LIAM R.- DOB: 01/06/1958- ISS: 01/06/2015- EXP: 08/12/2020- Address: 123 STREET ADDRESS, YOUR CITY WA 99999-1234- Sex: M- Height: 5'-08"- Eyes: BLU- Weight: 165 lb- Restrictions: B- Endorsement: L- Veteran: yes
Open full image
Scanned form
Proposal form with dense footer condensed
This noisy form mixes typed fields, checkboxes, handwritten marks, and a long legal footer.
Hard part: Noisy scan, checkboxes, handwritten signature
Instruction
Preserve the commercial form sections, but compress the long legal footer into a short note.
Base OCR Customize BBox
Markdown output
md
# Proposal 10675- Company: STOUT INDUSTRIES, INC.- Date: October 16, 1987- Customer: Lorillard Corporation- Address: 666 Fifth Avenue, New York, New York 10103- Attention: Mr. Robert Kennedy- Representative: Mr. A. D. Steinberg## Item- Item: Harley Davidson Metal Plaque- Size: 17 1/2 x 23 1/2- Gauge: .025- Colors: Transparent gold, opaque black, white and orange## Material and finish- Material: Aluminum- Base color: Aluminum- Single face: checked- Holes: yes- Number of holes: 4- Corners: square- Edges: hemmed- Stamp frame: yes- Embossed: yes## Commercial terms- Quantity: 500 plaques- Price: $9.18 each- One-time tooling: $3,015.00- Steel tips: $1,045.00- Billing: bill as manufacture- Warehousing: ship immediately- Terms: NET 10 DAYS## Notes- Price is based on reproduction of customer supplied "Pack" box.- Footer contains standard price-adjustment, freight, and liability language.
Open full image
Bill of lading
Blank bill of lading template as markdown
This sample is an unfilled bill of lading template, so the real OCR result is mostly section headers and placeholders.
Hard part: Logistics layout with party and shipment blocks
Base OCR Customize BBox
Markdown output
md
# Bill of Lading - Short Form - Not Negotiable- Page: 1 of 1- Bill of lading number: [blank]## Ship from- Name: [Name]- Street address: [Street Address]- City, state, ZIP: [City, ST ZIP Code]- SID No.: [blank]## Ship to- Name: [Name]- Street address: [Street Address]- City, state, ZIP: [City, ST ZIP Code]- CID No.: [blank]## Third party freight charges bill to- Name: [Name]- Street address: [Street Address]- City, state, ZIP: [City, ST ZIP Code]## Carrier details- Carrier name: [blank]- Trailer number: [blank]- Serial number(s): [blank]- SPAC: [blank]- Pro number: [blank]## Freight charge terms- Prepaid: unchecked- Collect: unchecked- 3rd Party: unchecked## Customer order information- Customer order no.: [blank]- Package rows: blank template- Additional shipper information: blank template## Signatures- Shipper signature/date: blank- Trailer loaded: by shipper / by driver- Freight counted: by shipper / by driver
Ready to test