ReportBuilder360
SECURITY OVERVIEW

Security you can explain.
To your auditors, too.

How ReportBuilder360 protects your data, your sign-ins and your reports, on your own servers or in a hosted organization. This page describes what the current version does.

WHERE IT RUNS

Your infrastructure.
Your choice of AI.

Where the application runs and where AI runs are separate decisions, and both are yours.

On your own servers

Install on infrastructure you operate, or start with a hosted organization. The same product either way.

A workspace per organization

Each organization has its own users, reports, uploads, data credentials, AI settings and backups.

Local or cloud AI

Run a model on your own Ollama host, or choose a supported cloud provider. With only a local model configured, nothing is sent to an outside AI service.

Figures computed first

ReportBuilder360 calculates comparisons, drivers and totals itself. AI only words the result, so it never invents a number.

AI changes wait for review

Every AI draft or change is shown for review before it becomes part of a report, and one Undo removes it.

Cloud AI only with consent

Designing from a PDF or photo uses a cloud model only after someone agrees to it for that file.

YOUR DATA

Read-only by design.
Narrowed to each reader.

Reports read your data. They never change it, and each person sees only the rows they are allowed to see.

Read-only queries

Every SQL query passes a read-only validator and uses named parameters, never text pasted into SQL. Write statements are refused.

Row rules on every result

Row-level security, written as people and allowed values, applies to reports, dashboards, analysis, AI answers and exports alike.

Encrypted credentials

Data-source passwords and keys use authenticated encryption with a separate key. A missing key stops access instead of guessing; screens never show saved secrets.

No calls into your network

REST sources, image imports and webhooks cannot reach internal network addresses unless the server operator allows a specific address.

Limits on every query

Results are capped at 10,000 rows per query and long-running queries time out, so one report cannot exhaust a database.

Encrypted in transit

HTTPS with HSTS, strict security headers and secure session cookies. Certificates can be managed from the application.

SIGN-IN AND ACCESS

Every sign-in option.
Off until you turn it on.

Administrators switch each option on per organization, so nothing changes for existing users until you decide.

Single sign-on

OpenID Connect and SAML 2.0 with your identity provider, such as Microsoft Entra ID, Okta or AD FS. SAML assertions must be signed.

Directory sign-in

LDAP and Active Directory, encrypted with LDAPS by default. A local administrator account always keeps working.

Two-step sign-in

Authenticator apps with recovery codes, required for everyone if you choose. Repeated failed attempts lock sign-in for 15 minutes.

Automatic provisioning

SCIM 2.0 creates, updates and deactivates people and groups. Deactivation keeps their reports and history.

Network allowlist

Accept sign-in, API, single sign-on, SCIM and embedded-report requests only from the networks you list.

Roles you can explain

Administrator, designer and reader, plus report permissions. Group mappings from your identity provider can only lower a role, never raise it.

INTEGRATIONS

Keys you can revoke.
In one click.

Connect business applications without handing out broad or permanent access.

Hashed application keys

Shown once, stored only as a hash, limited to named reports, and unable to run SQL or edit reports. Rotating a key revokes it and every session at once.

Short-lived embedded sessions

Your backend requests a session for an authorized user; it expires quickly and works only on the exact sites you list.

Signed webhooks

Every delivery carries an HMAC-SHA256 signature and a timestamp, with retries and a delivery log.

AUDIT AND RECOVERY

Know who did what.
Recover when you need to.

A clear record of activity, and the means to restore an organization or an earlier report.

Activity log

Sign-ins, changes, exports and administration are recorded with who did them and when.

Security alerts

Signals such as repeated failed sign-ins are surfaced for administrators to review.

Retention you set

Older events can be archived automatically. Archiving never deletes history.

Report history

Review changes to a report, compare versions and restore an earlier one.

Backups per organization

Scheduled and on-demand backups, downloadable for an off-server copy, with guarded restore.

Protected documents

PDF exports can carry watermarks, a password and a digital signature.

HONEST ANSWERS

Questions security teams ask.

Do you hold ISO 27001 or SOC 2 certification?

Not yet. During an evaluation we share this overview and architecture details, answer your security questionnaire, and you can run the product on your own servers to test it yourself.

Does AI see our data?

Only where you allow it. With a local model on your own hardware, nothing leaves your network for AI. With a cloud provider, ReportBuilder360 computes the figures first and sends the findings to be worded, through the provider and account you choose.

Can a report change our database?

No. Queries pass a read-only validator and run with named parameters. We also recommend a database account with read-only permissions, so the database enforces it too.

Where are backups kept?

On the server that runs ReportBuilder360. Administrators can schedule them, download a copy and keep it off the server, and restore an organization when needed.

How do we raise a security question or concern?

Contact us through the contact page and mention security; we will arrange a session with your security team.

MAKE YOUR NEXT REPORT A TURNING POINT

Bring the data.
We’ll explore the possibility.

Book a product conversation sales@reportbuilder360.com