marql

How marql handles your business data

marql reads your POS and accounting without changing them. Data stored in the EU on AWS. Encrypted in transit and at rest. Your data is never sold to third parties.

For compliance documentation, DPA agreements, or specific security requirements, contact .

Core security principles

Read-only on POS and accounting

marql connects to your POS and accounting with read-only credentials, so it cannot change a record, a price or a setting in them. If you run e:corg 2.0, e:corg itself creates a store order or transfer as its own draft once a person has approved it.

Encrypted in transit and at rest

All data in transit between your systems and marql is encrypted using TLS. Data stored in our platform is encrypted at rest using industry-standard encryption.

EU data storage

Your data is stored on AWS infrastructure in the EU: the database in Frankfurt (eu-central-1) and uploaded files in Stockholm (eu-north-1). AI inference runs at providers in the United States, so the data in each AI request leaves the EU while it is processed.

Isolated per organisation

Each organisation's data is logically isolated. Your data is not accessible to other organisations on the platform. A dedicated database is available, scoped separately, for physical isolation.

You control access

Role-based access means you decide who in your organisation sees which data. Owners see all stores; store managers see only their location. Access is configured once and enforced automatically.

No data sold, no ad tracking

We do not sell your data to third parties and we do not use advertising or tracking cookies. Session cookies are used strictly for authentication.

What marql can and cannot change

marql connects to your POS or accounting system with read-only credentials, usually an API key or export access that only allows viewing. Through these connections marql cannot:

  • Create, update, or delete transactions in your POS
  • Modify invoices, records, or entries in your accounting system
  • Change prices, products, or any configuration in your POS or accounting
  • Access data beyond what is needed to produce operational views

e:corg 2.0 is the only system that can take a document from marql. When a person approves a store order or transfer, e:corg creates it as its own draft. With other ERPs, your team gets the document as a file and enters it. Your POS and accounting keep working exactly as before. If you disconnect a source at any time, the read access is revoked immediately.

Who can give an AI employee authority

marql's AI employees give advice and prepare documents, such as a store order or a promo plan. What each of them may do is granted separately, under the rules below.

  • Authority is set separately for each type of action, on levels that run from Observes to Executes within limits.
  • Only the owner or COO can raise a level. They raise it one step at a time, and the reason stays in the history.
  • If a guardrail is breached on three readings in a row, the employee drops a level automatically. While data is stale, it can only advise.
  • Every action is logged with who or what did it.
  • Each employee can be paused.
  • The AI finance manager does not post to the books, make payments or change prices.
  • marql only reads POS and accounting data. The one system that can take a document from marql is e:corg 2.0, and e:corg creates that document itself after a person confirms it.
  • Demo organisations execute nothing.

The levels, and how each employee is measured against control stores, are on the AI employees page.

How your data flows through the platform

For transparency, here is what happens to your data at each stage of the marql pipeline:

Stage
What happens
Where it's stored
Connection
Read-only API key or scheduled CSV export from your POS / accounting system
Credentials stored encrypted, never logged in plain text
Sync
Transaction records, invoices, and metric data read from source systems
Normalised data stored in your organisation's isolated data store on AWS EU
Briefing
marql processes the synced data to produce daily briefings, anomaly alerts, and AI answers
Processed views stored in AWS EU; source raw data cached per your plan's retention window
Access
Users access briefings and dashboards through the platform
Sessions use TLS; role-based access limits what each user can view

For a full technical diagram of the data flow from your POS to the briefing, see the data flow page.

Infrastructure and hosting

marql runs on Railway in Amsterdam, with data stored on AWS in the EU (Frankfurt and Stockholm). Traffic passes through Cloudflare. AI inference runs at providers in the United States. We do not operate our own physical data centres.

Cloud provider

AWS (Amazon Web Services)

Data region

EU — Frankfurt (eu-central-1) for the database, Stockholm (eu-north-1) for files

Encryption in transit

TLS (all connections)

Encryption at rest

Industry-standard, enabled by default

Data transfers outside EU

AI inference requests, processed in the United States

Infrastructure provider (hosting)

Railway / AWS, data stored in EU

Dedicated database, scoped separately

By default your organisation's data is logically isolated — separated from others at the application and database level. For networks with specific data governance requirements, a dedicated PostgreSQL database or isolated infrastructure is available, where your data is physically separate from other organisations' data. It is scoped and quoted separately from the standard per-location price.

Scoped engagements can also include SLA agreements and custom data retention terms. For requirements like these, contact or see how large networks are scoped.

GDPR and your data rights

marql operates under GDPR for organisations and users in the European Economic Area. You have the right to access, correct, or delete your personal data at any time.

  • Request a copy of the personal data we hold about you
  • Request correction of inaccurate personal data
  • Request deletion of your account and associated data
  • Object to processing or request restriction
  • Request a Data Processing Agreement (DPA) for your organisation

To exercise any of these rights or to request a DPA, contact . Full details are in the Privacy Policy.

Third-party services

We use a limited set of sub-processors. The following third-party services may handle data as part of the marql platform:

Service
Purpose
AWS / Railway
Infrastructure hosting — database in Frankfurt, files in Stockholm, application in Amsterdam
Fireworks AI / Together AI / Groq
AI inference in the United States
Google / Microsoft OAuth
Authentication only — name and email received
Sentry
Error monitoring — anonymised stack traces only
Google Analytics 4
Website analytics — only with explicit consent
Calendly, Inc.
Demo scheduling — name and email for meeting booking only

Full list: https://platform.marql.one/subprocessors

We do not sell personal data or business data to any third party. We do not use advertising networks.

Security questions

DPA, compliance documentation, or specific requirements?

For Data Processing Agreements, penetration test reports, security questionnaires, or enterprise compliance discussions, contact our privacy team directly.

Read the Privacy Policy

Frequently asked questions

Data security and privacy

Your data is stored on AWS infrastructure in the EU: the database in Frankfurt (eu-central-1) and uploaded files in Stockholm (eu-north-1). AI inference runs at providers in the United States, so the data in each AI request leaves the EU while it is processed.

No. marql reads your POS and accounting and never creates, modifies or deletes records in them. If you run e:corg 2.0, e:corg creates an approved store order or transfer as its own draft after a person confirms it. Other ERPs receive the document as a file.

Yes. All data in transit between your systems and marql is encrypted using TLS. Data stored in the platform is encrypted at rest using industry-standard encryption.

Yes. Contact to request a DPA for your organisation. Custom data governance terms are available for networks that are scoped individually.

Disconnecting a source revokes marql's read access immediately. You may request deletion of your account and all associated data at any time by contacting .