DynamicForm Start free trial
Step 14 of 15 · Architecture and security · topic 1 of 4

Trust · Architecture and security

Built to keep your data and your rules safe

Designers write configuration; the engines read it. Every rule the screen shows is checked again on the server.

Every part of the platform behind guarded gatesA small city of four districts — designers, render engine, workflow engine and data stores — joined by roads with shield gates at each crossing.DesignersRender engineWorkflow engineData storesGuarded gates: server-side checks, roles, encrypted sign-in

Rules enforced twice

Screen and step access rules are applied in the browser and re-checked on the server for every submission.

Roles and branches

Menus, cases and fields limited by role, branch and process step.

Full history

Every workflow move is recorded with who, when and the data at that moment.

Encrypted sign-in

Encrypted credentials, two-factor codes, session revoke, idle-time warning.

Your choice of database

Transaction data in normalised tables on SQL Server, Oracle or MySQL; case history in a document store.

Keys stay on the server

Outside services are called by the server; phones and browsers never see credentials.

In short

Transaction data sits in third-normal-form tables on SQL Server, Oracle or MySQL, and the platform serves web and phone over APIs. Rules shown on screen are checked again on the server.

Connected through the AI Integration Engine

Vault-backed credentials

A vault is protected storage for passwords, keys and other service credentials. Production uses OpenBao, and values are entered write-only so the secret is not shown back after entry. The server makes provider calls, keeping those credentials off browsers and phones.

Mask and retain

Masking hides personal information before it appears in a call trace or an AI request. A trace is the record of what happened during a service call, used to investigate problems. Masked traces are kept for 90 days by default, with retention adjustable within supported limits.

Protect company boundaries

Company isolation keeps each customer’s data separate. Connections to internal network addresses are blocked unless the destination is explicitly permitted. TLS 1.2 or higher protects data in transit, while response-size limits and safe XML reading control how provider replies are received.

Explore integrations

Keep every company separate

Each company has its own users, applications and data, with separation across every data layer.

Explore multi-company SaaS · Explore security and identity

Separate stores for each application

Place application data where it needs to live within platform policy. Records, documents, files, logs and cache can use company stores or separate application stores. Files stay only in the company's or application's own file store.

Choose logical separation within a shared store or physical separation in an application's own database, including on another server. Move layers when needed, including back to company stores. Only the affected application pauses during the move, and its logs follow it.

My tasks and dashboards still gather that application's work across stores. Application boundaries stay clear while the storage arrangement changes. Explore Applications for the storage choices and the company administrator's controls.

Application storage diagram compares company, logical and physical stores across records, documents, files, logs and cache.
Choose each application's stores by data layer within platform policy.

Checked before release

Review changes before they reach end users. Every change receives a security review, then a correctness review before release.

The latest checks include more than 3,500 automated tests, company-separation checks, mobile app tests and real-browser walk-throughs. The checks completed with no failures, and company-separation sweeps found no leaks.

Application boundaries also guide the design: shared company information stays separate from application work, and stores follow the application they serve. Explore platform operations for the controls used to run and monitor the service after release.

Platform diagram contains separate company areas, each holding application cards behind clear company and application boundaries.
Applications sit inside companies, with separate boundaries at each level.

Follow service health

Give platform operators a view of service health. The server monitor shows CPU and memory per server, plus database, document store and cache health. It also shows the health of each company's separate stores.

The monitor keeps 24-hour history, helping operators compare current readings with recent activity. When a limit stays crossed for several minutes, it raises an alert in the platform log. Current cards and history support review of what changed.

These controls sit in the separate operator console. Explore Platform operations for operator sign-in, permitted console sections and the recorded sessions used when working inside a company.

Build your first app free

All platform features, hosted by us for 120 days. For testing and evaluation, subject to our Fair Use Policy.

Updated 16 November 2026