FormParse
Schema-first document extraction

Documents in. Structured data out.

Turn invoices, receipts, and business PDFs into validated JSON, CSV, or Excel in minutes. Or automate everything over a typed REST API.

Start for freeView API docs

Free plan · 10 pages/month · no credit card

Workflow

Four steps from file to dataset

The same pipeline behind the portal and the API — nothing bespoke, nothing manual.

01

Define a schema

Six fixed templates — invoice, receipt, purchase order, delivery note, contract, generic — or your own custom fields with types and required flags.

{} invoice · 7 fields
02

Upload documents

Single files or batches of up to 100 — PDF, JPG, PNG, HEIC. Page counts are computed at intake.

↑ batch · 24 files
03

Review extracted values

Field-level confidence and validation errors flag anything that needs a human look before it ships.

needs review
04

Export or pull via API

Download JSON, CSV, or Excel — or poll jobs and fetch results straight from /v1.

GET /v1/extractions/{id}

Capabilities

Built like data infrastructure

Every feature exists to get documents into your systems as clean, predictable records.

Batch processing

Up to 100 documents fan out through per-organization queues, run asynchronously, and land in one aggregate dataset.

doc_014.pdf
completed
doc_015.pdf
processing
doc_016.pdf
queued

Review & validation

needs_review status, field-level confidence, and explicit validation errors — nothing fails silently.

due_dateneeds review
total_amount 0.97

validation_errors: ["due_date is required"]

Schema-first extraction

Results always match the contract you defined — hover a field to see where it lands.

supplier_name→ result.data.supplier_name.value
invoice_number→ result.data.invoice_number.value
total_amount→ result.data.total_amount.value

JSON, CSV & Excel export

Per document or aggregated per batch — one clean dataset, ready for your pipeline or your accountant.

.json.csv.xlsx

REST API + OpenAPI 3.1

The spec is generated from the same Zod contracts the app runs on, with an interactive Scalar reference.

GET /v1/openapi.json

Retention controls

Pick a window per organization. A daily cleanup deletes expired data and writes audit logs.

0d7d30d90d

Usage-based billing

Hard monthly page quotas, metered at intake. Platform errors never consume your quota.

1,540 / 2,500 pages62%

API keys, done right

SHA-256 hashed at rest, shown exactly once, revocable instantly. Key management stays session-only.

dk_live_ ••••••••••••

Developers

An API you would have built yourself

One resource model across portal and API. Sync for small documents, 202 + polling for async jobs and batches.

  • Bearer auth with dk_live_ keys — hashed at rest, shown exactly once
  • OpenAPI 3.1 generated from shared Zod schemas, Scalar reference at /docs
  • Typed TypeScript SDK with response validation and rich errors
  • Rate limits with Retry-After and X-RateLimit-* headers
View API docsOpenAPI JSON
POST /v1/extractions
$ curl -X POST "$APP_URL/v1/extractions" \
-H "Authorization: Bearer dk_live_…" \
-H "Content-Type: application/json" \
-d '{"document_id": "9b2f…", "async": false }'
200 OK · application/json
{
"job": { "id": "5f0c…", "status": "completed", "mode": "sync_live", "page_count": 1 },
"result": {
"data": {
"supplier_name": { "value": "Example Supplies AG", "confidence": 0.98 },
"invoice_number": { "value": "INV-2041", "confidence": 0.99 },
"total_amount": { "value": 1240.0, "confidence": 0.97 }
},
"validation_errors": []
}
}

Data protection

Your documents, on a timer

Extraction needs your most sensitive files. The platform is built so they do not stick around.

Retention on a timer

0, 7, 30, or 90-day windows per organization. A daily job deletes expired documents, results, and exports — storage objects and rows.

Audit trail

Deletions and credential lifecycle events write audit logs: document.retention_deleted, api_key.revoked, and more.

Scoped access

Postgres row-level security isolates every organization. API keys are SHA-256 hashed, shown once, and revocable instantly.

EU processing posture

Queue and rate-limit infrastructure run in EU regions (Trigger.dev, Upstash). An EU-only extraction mode is planned.

Read the full security overview

Pricing

Quota-based. No surprises.

Pages are counted at intake; only billable processing consumes quota. Manage everything from the billing page.

Free

Free

10 pages/mo

10 pages per month for evaluation and light usage.

  • Schema-first extraction
  • JSON, CSV & Excel exports
  • Portal access only (no API)
Get started
1 month free

Casual

CHF 12.90/mo

200 pages/mo

200 pages per month for regular personal use.

  • Everything in Free
  • 200 pages per month
  • Self-serve billing in CHF, EUR, or USD
Start with Casual
API access

Pro

CHF 39.90/mo

2,500 pages/mo

2,500 pages per month with full API access.

  • Everything in Casual
  • Full REST API access
  • Manage or cancel anytime from the portal
Start with Pro

Enterprise

Custom

Unlimited pages

Custom page quota and contract billing. Contact us.

  • Everything in Pro
  • Custom page quota
  • Contract billing
Contact sales

Prices shown in CHF. EUR and USD available at checkout. Monthly plans include a one-time 1-month free trial. Card required; cancel before the trial ends to avoid charges. Manage billing

Start turning documents into datasets

10 free pages every month. The first extraction is one curl away.

Start for freeView API docs