HyperAPI
DashboardDocsBlogsPricing
Sign inContact us
HyperAPI
by Hyperbots
HyperAPI
DashboardDocsBlogsPricing
Sign inContact us

Financial document intelligence · by Hyperbots

Turn any document
into structured data.

Documents go in. Clean, validated JSON comes out. Parse, extract, classify, split, and redact in one call your code and agents can trust.

Start freeTry the APIsRead the docs

raw documents in · structured data out

Works with your agents
Claude CodeCodexOpenCodeCursorGitHub Copilot
Build it yourself

Every document pipeline breaks.

OCR, rules, models, glue — HyperAPI is that pipeline, behind one call.

Misreads totalsOCR · TOTAL = 508.20Breaks per vendorregex ✓ "INV-88241"Can’t decideinvoice .54 · rcpt .46Brittle mappingrow[3].cols[2] → totalOne miss = leakpay J.D · acct 4417…Pages you at 3am429 · retry 3/5Rebuilt, againfinal_v2(3).py · ×5
one endpoint

POST /v1/extract

{ "document_key": "inv_88241.pdf" }

invoice_numberINV-88241total508.20 USDline_items4
200 OKvalidated JSON$0.08 / page
parseextractclassifysplitredact

All of it, replaced by one API.

Operations

What one call does to your document.

Each operation reads the page, understands it, and returns what you need.

Select an operationcycling
document.pdf3 pages · $508.20Try it now
Get started

Add HyperAPI to your app in minutes.

Point the Python SDK at a file and call any operation. No upload step. Pick one on the right and the code updates; the cURL tab shows the raw REST flow.

QuickstartPOST/v1/extract
from hyperapi import HyperAPIClient
client = HyperAPIClient(api_key="hk_live_…")

# the SDK takes a file path — no upload step

fields = client.extract("invoice.pdf")
print(fields)
Operations
Parse/v1/parse
Clean markdown, layout preserved.
Extract/v1/extract
Typed JSON fields and line items.
Classify/v1/classify
Identify the document type.
Split/v1/split
Separate a bundled PDF.
Redact/v1/redact
Mask PII before the model.
Get your API key
REST API/v1 referencePython SDKpip install hyperapiMCPfor AI agentsPlaygroundtry in browser
For agents

It compiles your goal into a pipeline.

Coding agents already speak plain English and write code. What they can’t do is read a messy document. So HyperAPI hands them that, as five operations they compile into the exact pipeline your goal needs.

The agentlanguage + code
Accounts payable

“Reconcile this month’s vendor invoices against our open POs, flag anything over $500, and mask the bank details before it moves on.”

picks and orders the operations it needs
318 vendor invoices · PDF + scans
compiles to
HyperAPIdocument operations
[ compiled ]4 ops
classifyinvoice vs. the rest
splitburst the batch → one doc each
extractvendor · total · line items
match→ open purchase orderagent
flagtotal over $500agent
redactmask the bank details
One reconciled ledgermatched · flagged · per-field citedone JSON out
+2 agent steps

the agent’s language + code × HyperAPI’s five document operations = one cited, structured result

Join the Agents beta
???{ }</>

FAQ

Questions, answered.

Everything worth knowing before your first call. Tap a question to expand it.

Still stuck?
We usually reply within a few hours, on a real engineer's keyboard.
Email us →Read the docs
Five operations against any file: parse to clean markdown, extract to typed JSON fields, classify to route by type, split to separate bundled documents, and redact to mask sensitive data. Upload once, get a key, and call any endpoint against it.
parseextractclassifysplitredact
PDFs, scans, and images (PNG, JPG) across 30+ formats. Advanced parsing handles multi-column layouts, tables, and low-quality scans with layout-aware OCR.
Every response is schema-validated JSON with per-field confidence scores and source bounding boxes, so you can verify exactly where each value came from. Pick Fast, Balanced, or Thorough depending on your accuracy and latency needs.
Files are processed in memory and deleted the moment a run completes, unless you explicitly enable retention. Nothing is used for training. See the security page for our data-handling and security details.
Yes. The Agents beta lets you hand over a stack of documents and a goal; the agent picks the operations, runs them in sequence with guardrails, and returns one structured result. Drive it from Claude Code, Cursor, or any MCP client.
Pay per page, priced by operation. New accounts get 10,000 pages free every month, no credit card required. Volume and enterprise plans are available, see pricing for the full breakdown.
An official Python SDK, plus a plain REST API that works from any language with cURL. Webhooks and async jobs are built in for bulk and long-running files.
PDF
01 · 02 · 03 · 04 · 05
hyperapi · document intelligence · structured data · invoice · receipt · contract
HyperAPI

Start with a free account.

No credit card. 10,000 pages free every month. Read the docs and ship today.

Start free →
$ pip install hyperapi
HyperAPI

Document intelligence APIs from Hyperbots. Built for dashboard users, backend services, and AI agents that need reliable document output.

Product

  • Dashboard
  • UI Sandbox
  • Pricing
  • Contact Sales

Docs

  • Quickstart
  • API Reference
  • Upload API
  • Async Jobs
  • SDKs
  • Status

Company

  • Hyperbots↗
  • About↗
  • Blog
  • Research↗

Legal

  • Privacy Policy↗
  • Terms and Conditions↗

© 2026 Hyperbots Inc. All rights reserved.

Privacy Policy·Terms·hyperapi-support@hyperbots.com