For startups

Ship document extraction in hours. Not months.

Q: How fast can we go live?

First API request in under an hour. First production extraction within a week, including auth wiring, error handling, and webhook callbacks. Most startups ship their MVP integration in a sprint.

Q: What's the actual price?

Pay-per-page from $0.005-0.02 depending on document complexity. Starter plan is $79/mo on annual or $99/mo monthly, includes 1,000 documents and six document types. Above that it's per-page. No setup fees, no implementation services.

Q: Do you offer a free tier?

We offer a $200 trial credit on signup — enough for ~10,000-40,000 documents depending on type. Use it on real production traffic. Cards on file aren't charged until the credit is exhausted.

Q: When should we move to Enterprise?

When you start losing deals because procurement asks for SOC 2 reports under NDA, single-tenant infrastructure, or a custom DPA. The transition is typically painless: same API, same schema, just a different account tier.

Q: Will you train on our data?

No — never. Customer data is not used for model training, evaluation, or service improvement. This is contractually enforced in every plan, including Starter.

Building your own OCR pipeline costs 2-3 engineering months and still breaks on real-world PDFs. fluex gives you production-grade extraction via a REST API you can call in an afternoon — no ML team required.

Start for free Read the docs

The challenge

Why building it yourself doesn't work.

Engineering bandwidth

Building custom parsers burns cycles that should go to core product. And bespoke parsers break every time a supplier changes their template.

Real-world PDF variability

Production documents are rotated, low-quality, multi-column, or handwritten. Regex and template-based approaches break the moment you encounter a new format.

Prototype-to-production pricing

Tools that are cheap to start become expensive at scale — or require enterprise contracts before you've found product-market fit.

How fluex helps

One API call. Clean JSON.

fluex abstracts the entire document processing stack. You upload a file, we return structured data. No configuration, no ML training, no template maintenance.

REST API + 4 official SDKs — C#, Python, Node.js, and Go. Working integration in under 30 minutes.
Pay-per-page, no minimum — start processing your first document for free. Scale to millions of pages without renegotiating a contract.
Template-free extraction — models adapt to layout changes without retraining. Works on documents you've never seen before.
Webhooks & async jobs — don't block your API with polling. Subscribe to extraction events and react when documents are ready.

extract.py

# fluex SDK — Python
import fluex

client = fluex.FluexClient("fx_live_...")

result = client.documents.extract(
  type="payslip",
  file="payslip.pdf"
)

# {
#   "gross_salary": 12480,
#   "net_salary": 9856,
#   "employer": "Acme Ltda.",
#   "reference_month": "2026-03",
#   "confidence": 0.97
# }

<30m

From sign-up to first extraction

Official SDKs (C#, Python, Node, Go)

Minimum monthly commitment

99.2%

Accuracy on structured documents

Startup-friendly

Ship document AI in a week.

Startups don't have the time or the budget for a six-month integration. fluex's Starter and Growth plans are designed for the 1-5 engineer team that needs extraction in production this quarter — pay-per-page pricing, no annual commit, first integration shippable in days, not months.

No annual commit

Month-to-month billing on usage-based plans. Cancel anytime. Pause usage between fundraising rounds without renegotiating.

First integration in < 1 week

Six pre-trained document types live on Starter. Drop in the API key, send your first request in under an hour.

Founder-friendly contracting

Standard MSA, standard DPA, e-signature. No procurement gauntlet. We've signed contracts with Series-A startups in under three days.

Embed-in-product, not redirect

All extraction flows are headless API calls. Your users never see a third-party portal. Your branding stays intact.

Stack patterns

Drops into the YC stack.

fluex ships SDKs and pre-built patterns for the toolchain most modern startups already run. No custom middleware, no "call your sales rep" integrations.

Backend SDKs — First-class clients for Python, Node.js, Ruby, Go, and .NET. Type-safe responses, idiomatic error handling, automatic retries with backoff.
Frontend — React Native and Flutter components for in-app document capture with mobile-optimised compression and on-device validation.
Workflow — Native integrations with n8n, Zapier, Make, and Temporal. Most startups ship their first workflow without writing backend orchestration.
Auth — Auth0, Clerk, WorkOS — SSO and SCIM ready when you start landing enterprise pilots.
Storage — S3-compatible storage hooks. fluex never holds your documents longer than your retention policy says it should.

FAQ

Common questions.

How fast can we go live?

First API request in under an hour. First production extraction within a week, including auth wiring, error handling, and webhook callbacks. Most startups ship their MVP integration in a sprint.

What's the actual price?

Pay-per-page from $0.005-0.02 depending on document complexity. Starter plan is $79/mo on annual or $99/mo monthly, includes 1,000 documents and six document types. Above that it's per-page. No setup fees, no implementation services.

Do you offer a free tier?

We offer a $200 trial credit on signup — enough for ~10,000-40,000 documents depending on type. Use it on real production traffic. Cards on file aren't charged until the credit is exhausted.

When should we move to Enterprise?

When you start losing deals because procurement asks for SOC 2 reports under NDA, single-tenant infrastructure, or a custom DPA. The transition is typically painless: same API, same schema, just a different account tier.

Will you train on our data?

No — never. Customer data is not used for model training, evaluation, or service improvement. This is contractually enforced in every plan, including Starter.

Start building

Your first extraction in 30 minutes.

Get your API key Read the docs