Submit a PDF, get back typed elements — text, equations, figures, tables — each with pixel-precise bounding boxes. One endpoint.
See the output
Upload a real PDF. Get typed elements with bounding boxes.
curl -X POST https://api.bitparse.ai/parse \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: multipart/form-data" \ -F "file=@document.pdf"
{
"elements": [
{
"id": "page1_elem0",
"page": 1,
"element_type": "sub_title",
"bbox": {
"x1": 223,
"y1": 801,
"x2": 436,
"y2": 829
},
"content": "## 1. Introduction"
},
{
"id": "page1_elem1",
"page": 1,
"element_type": "text",
"bbox": {
"x1": 221,
"y1": 844,
"x2": 1050,
"y2": 1136
},
"content": "Commerce on the Internet has come to rely almost exclusively on financial institutions serving as tr..."
},
{
"id": "page1_elem2",
"page": 1,
"element_type": "text",
"bbox": {
"x1": 221,
"y1": 1136,
"x2": 1050,
"y2": 1326
},
"content": "What is needed is an electronic payment system based on cryptographic proof instead of trust, allowi..."
}
]
}
What you get back
Typed Elements
You get back text, headings, equations, figures, tables, and images — each classified by type. Not raw strings.
Bounding Boxes
Each element includes normalized coordinates [x1, y1, x2, y2]. Crop regions, reconstruct layouts, or build spatial indexes.
Structured JSON
You get typed JSON responses — no regex cleanup, no post-processing, no second pass.
Parallel Processing
Your pages process in parallel. Multi-page PDFs don't mean multi-minute waits.
One POST Endpoint
POST /parse with a PDF, PNG, or JPEG. That's the entire API. No config objects, no pipeline stages, no setup.
API Key Auth
Authenticate with a single header. Your keys are SHA-256 hashed at rest and revocable instantly.
$0.008 per page. No subscription
1 credit = 1 page. Credits never expire. Buy more anytime.