Do you train on our data?

No. Customer data is never used to train models. The base models are pre-trained and customers' inputs and outputs are isolated. We can sign a DPA confirming this in writing for procurement.

Can it handle a 200-page KYC pack?

Yes. anyformat handles long, mixed-document packs: ID, proof of address, payslips, bank statements, source of funds attestations. It classifies each section and extracts the right fields, even when documents are merged into a single PDF.

Does it integrate with our core banking and risk engine?

Yes. REST API or webhooks into Temenos, Mambu, Finacle and in-house cores; Provenir, FICO, GiniMachine for risk; LexisNexis, ComplyAdvantage for screening. Extracted fields land in the right place without a middle layer.

What about PEP and sanctions screening?

anyformat extracts the data; we don't replace your screening tools. The extracted name, DOB, nationality and address fields are passed straight to your existing screening vendor with high accuracy, so false positives from typos disappear.

Is there an on-premise option?

Yes. For tier-1 banks and insurers with strict data residency or air-gapped requirements, anyformat is available as a private deployment on your VPC or on-premise, with the same accuracy as the cloud product.

Yourops teamshouldn't bea data entryteam.

Every KYC pack, payslip, bank statement and policy your customers send still gets reviewed by hand. anyformat reads them across formats and languages, scores confidence per field and pushes structured data into your core systems, with EU sovereignty and zero retention.

Book a demo

Financial operations team reviewing KYC and policy documents

ISO 27001 certified|EU sovereignty, zero retention|99% extraction accuracy

KYC PACK

STATEMENT

PAYSLIP

Core banking

Risk engine

Compliance

THE PROBLEM

Financial documents arrivein every format imaginable,and your team reviews them by hand.

Banks, fintechs and insurers run on documents customers send from their phone, their employer, their bank. PDFs, scans, screenshots, mobile photos, in different layouts and languages. Each one a compliance risk if it's mis-read.

Risk

missed field on a KYC pack is a compliance incident

ID expiry dates. Proof of address date. Source of funds amount. A miss isn't a typo, it's a regulatory finding when the regulator audits the file. Manual review can't prove it caught everything.

Volume

10x

spike during onboarding campaigns

A consumer lending or insurance promo turns a steady review queue into a backlog overnight. Compliance teams can't hire 10x analysts in a week. The result: slower onboarding, lost customers, queues that drag into next quarter.

Sovereignty

data residency is non-negotiable

Customer PII can't leave the EU. Many AI tools route through US infrastructure or train on customer data by default. anyformat is EU-native, ISO 27001 certified, with zero-retention processing and contractual GDPR alignment.

Time-to-decision

48h

manual KYC review, target is under an hour

Time-to-yes is a competitive metric. Manual KYC review chains a customer's onboarding to a human's queue. Automating the extraction step alone collapses 80% of that time and frees analysts for the cases that actually need them.

Trusted in regulated finance by

Built for financial documents

Trained on the documentsyour ops and compliance teamsactually receive.

Not generic PDFs. The specific artifacts that flow through onboarding, lending and underwriting, with the sensitivity those documents require.

Bank statements

Period	Bank	IBAN	Inflows	Outflows
Apr 2026	Santander ES	ES84 0049···2104	4,820 €	4,210 €
Apr 2026	BBVA ES	ES79 0182···7710	2,415 €	2,180 €
Apr 2026	CaixaBank	ES19 2100···8841	6,120 €	5,950 €
Apr 2026	Sabadell	ES12 0081···3304	1,840 €	1,720 €
Apr 2026	Bankinter	ES40 0128···5612	3,290 €	3,050 €
Apr 2026	ING ES	ES61 1465···0298	2,680 €	2,510 €

Security and certifications

Enterprise-grade security

anyformat meets enterprise security and compliance requirements out of the box, so teams can run document workflows in production with confidence.

GDPRPersonal data handled to GDPR standards, with EU data residency.

ISO 27001Independently certified to ISO 27001 for information security management.

Self-hosted deploymentRun anyformat in our cloud, in a dedicated VPC, or fully self-hosted in your own infrastructure. Your data stays exactly where your policy requires.

Visit the Trust Center

Built for the agentic era

By engineers and agents
for engineers and agents

Call anyformat workflows straight from your codebase with the SDK or CLI, or drop in the agent skill and let your coding agent build and run them for you in minutes, not months.

Terminal

Claude Code

Opus 4.8

SDKs

Install the SDK to call anyformat workflows straight from your codebase.

npm install @anyformat/sdk

CLI

Use the Python CLI to run and manage your workflows from the terminal.

uv tool install --python 3.13 anyformat && afx --help

Agent Skill

Add this skill to teach your agent the anyformat workflow API.

npx @anyformat/skill

Beyond the model

Everything around the model, built in.

Advanced OCR and parsing, enhanced by the capabilities required to build and run reliable workflows, fast and at scale.

annie builds your workflow

annie configures the whole workflow for you in seconds. No manual setup.

I want to extract info from invoices

Field-level confidence

Every field gets a calibrated score: only review what needs it.

Contract name

99%

Software License Agreement

Date

80%

2026-03-24

Contract type

98%

Employment

Evaluate and monitor

Eval every change against your own dataset before it ships. Once live, track accuracy field by field.

Select the right mode

Switch between Fast, Standard, and Agentic. Optimize for latency or accuracy, depending on what each job needs.

Fast

Standard

Agentic

annie keeps improving it

Ask annie to improve your workflow. She finds the fix and applies it automatically.

Accuracy by iteration

1st Optimization

5 fields · 6 iterations

Completed

92.5%97.0%+4.5p

Why anyformat

Built for production,
not demos

Raw OCR can't read structure. A raw LLM hallucinates and can't be audited. anyformat combines both with deterministic rules, confidence scoring and validation, so the output is structured, traceable and ready for production.

Traditional OCR

Raw LLM

anyformat

Reads layout, tables and structure

–

Long documents, hundreds of pages

–

Complex, multipage tables

–

Poor scans and low-quality inputs

–

Any format and language, no templates

–

Visual citations: every value cited to its exact position on the page

–

Confidence scoring and human-in-the-loop

–

EU data residency, ISO 27001, zero retention

–

Questions about financial document AI

How anyformat fits into a regulated finance operation.

Inside the EU. anyformat is EU-native and runs on EU-hosted infrastructure. Zero retention by default: documents are processed in memory and not stored after the response is returned. Contractual GDPR alignment, ISO 27001 certified.

Start with your hardest documents.

anyformat does the heavy lifting on the documents that break other tools. Parse, extract and validate them into clean, reliable data, and get to production in minutes.

No credit card required · 50,000 free credits to start