Security First Architecture

Trust is our Operating System.

Sync-o is built on the Atlassian Forge platform, inheriting world-class security, compliance, and isolation by design.

Encryption at Rest & Transit
GDPR Compliant Processing
Zero Training AI
Security Architecture AI Controls Data Privacy Subprocessors Contact

Security Architecture

Defense in depth, leveraging the Atlassian Shared Responsibility Model.

Platform Isolation

Sync-o runs within the Atlassian Forge sandbox. We utilize our own secure AWS infrastructure for specialized processing and storage, ensuring strict isolation and control.

View Forge Security Docs

No Credential Access

We never see your passwords. Sync-o operates only within the OAuth 2.0 scopes your administrator authorizes at install time, scoped to the Atlassian products (Jira and Confluence) where the app is installed.

Enterprise Encryption

In Transit: TLS 1.2+ (TLS 1.3 where supported by client).
At Rest: All persisted data — configuration, content excerpts, embeddings, audit logs — encrypted with AES-256 via AWS KMS. Customer AI keys are additionally protected via Forge setSecret.

AI Safety & Controls

You Control the Intelligence

We treat Generative AI as a "stateless" processor. Your intellectual property is never used to train our models.

Zero Training Policy

By default, Sync-o routes all inference through Google Vertex AI (Gemini) in Belgium. Vertex AI's enterprise contract prohibits using customer prompts or responses to train Google's models. BYOM providers (OpenAI, Anthropic, Azure OpenAI) operate under their own commercial API contracts, which also exclude API data from training by default.

Version History Safety Net

Every Sync-o update is published with full Confluence version history — review exactly what changed and revert with one click if needed.

Bring Your Own Model

Advanced plan users can connect their own private Google Gemini (via Regular API or Vertex AI), OpenAI, Anthropic, or Azure OpenAI accounts.

Data Handling

Sync-o is designed to minimize the data we store. We persist configuration required to operate the app, and we log limited operational metadata needed for reliability and support.

Data Retention

Full ticket / page bodies (for AI generation) In-memory only, never persisted
Content excerpts & vector embeddings (Smart Picker) 90-day TTL, deleted on uninstall
Configuration & encrypted credentials Deleted immediately on uninstall
System logs (CloudWatch) 90 days

For full details, please review our Privacy Policy and Data Processing Addendum.

Subprocessors

We rely on vetted third parties to deliver the service. Locations are selected to maximize GDPR compliance.

Entity Purpose Location Engagement
Amazon Web Services Cloud hosting, data storage, serverless compute (Lambda, DynamoDB, SQS, CloudWatch) Ireland (eu-west-1) Always
Google LLC (Vertex AI) AI content generation + embeddings (Gemini, text-embedding-004) Belgium (europe-west1) Default AI provider
OpenAI, OpCo, LLC AI content generation United States Optional — when Controller selects OpenAI in BYOM
Anthropic PBC AI content generation United States Optional — when Controller selects Anthropic Claude in BYOM
Microsoft Corporation (Azure OpenAI) AI content generation United States (legal entity); Controller-selected Azure region for processing Optional — when Controller selects Azure OpenAI in BYOM

Atlassian Forge is the underlying platform on which Sync-o is deployed; customers contract with Atlassian directly for the Forge platform, so Atlassian is not listed as a Sync-o sub-processor here. Operational tools that do not process customer Atlassian content (e.g., GitHub for source control, Cloudflare Pages for marketing-site hosting) are likewise excluded from this list.

Vulnerability Disclosure

Security researchers play a vital role in keeping our customers safe. If you believe you've found a vulnerability, please let us know immediately.

CSA CAIQ Lite Self-Assessment

Our published response to the Cloud Security Alliance's 124-question Consensus Assessments Initiative Questionnaire (Lite, v4.0.3) — covering encryption, IAM, change management, data handling, incident response, and 13 other security domains. Effective May 11, 2026.