Security & compliance

Built for European enterprise from the first line of code.

Your data matters. We treat it that way: encrypted, EU-hosted, traceable, deletable on request.

Contact the security team See sub-processors

Security

End-to-end encrypted. Continuously audited.

Your data is encrypted in transit (TLS 1.3) and at rest (AES-256). OAuth tokens are encrypted column-by-column via pgcrypto. Immutable audit log for every action.

TLS 1.3 on all communications
AES-256 at rest (PostgreSQL + S3)
OAuth tokens encrypted via pgcrypto
Append-only audit log + S3 WORM copy
Automatic secret rotation (Doppler)
MFA required for admins

Compliance

Native GDPR. SOC 2 in progress.

Built for European enterprise from the first line of code. GDPR, records of processing, DPA ready, right-to-be-forgotten implemented.

GDPR: records of processing up to date
DPA signable online (Enterprise template)
Right to be forgotten: cascade purge per tenant
Configurable retention policy
SOC 2 Type 1: Vanta audit started, delivery M9
ISO 27001: targeting Y2

Sovereignty

Hosted in France. EU sub-processors.

All your data stays in the European Union. Scaleway (Paris) or OVHcloud hosting, EU-managed databases, GDPR-compliant sub-processors.

Scaleway Paris hosting (PAR1/PAR2)
Aiven managed databases (Frankfurt / Paris)
Anthropic API: EU residency as soon as available
No transfer outside the EU without your explicit consent
Sub-processors: up-to-date public list

Isolation & access

Tenant isolation. Granular access.

Your data never crosses another customer's — isolation is enforced at 2 levels: application filter (Hibernate) AND Row-Level Security (RLS) policy directly in Postgres. No application code can read another tenant.

Postgres Row-Level Security: DB-level isolation
Tenant roles: Owner / Admin / Manager / Member
Project roles: LEAD (gets alerts) / MEMBER (read)
Read-only OAuth on Jira / GitHub / GitLab / Bitbucket / Confluence
No write to your tools without user confirmation (V2)
Per-project quiet mode (DORMANT) — immediate PM kill-switch

Data handling

Strict minimum. Always traceable.

We only collect what's directly useful to analyze your delivery. Nothing more.

Data collected

Delivery metadata only: tickets, statuses, assignees, messages, commits, PRs. Not your source code, not attachments, not personal data beyond professional identity.

Data sent to LLM

Only fragments needed for the requested analysis. No source code sent. Meeting transcripts are anonymized before analysis. 'On-premise LLM' mode (vLLM/RunPod) available in Enterprise.

Your keys, your rules

OAuth only, no password storage. You can revoke access at any time from your Jira/GitHub/Microsoft admin. Full data export available on request.

Sub-processors

Public list. Real-time updates.

You have the right to know who touches your data. Here's the exhaustive list of our sub-processors — modified only with 30-day notice.

Sub-processor	Purpose	Region
Scaleway	Compute + storage hosting	France (PAR1/PAR2)
Aiven	Managed PostgreSQL + Kafka	EU (Frankfurt / Paris)
Anthropic	LLM (Claude API)	EU residency enablable
RunPod	GPU for self-hosted models (option)	EU (NL / DE)
Microsoft Azure AD	OAuth Teams / Graph API	EU
Atlassian	OAuth Jira (outbound only)	EU instance possible
GitHub	OAuth GitHub (outbound only)	Depends on your instance
Doppler	Secret management	US (client-side encryption)
Grafana Cloud	Metrics + logs + traces	EU

Last update: 2026-06-19. Email notification for any addition / removal.

Trust center

Documents available on request.

For security teams that need more.

DPA template (PDF)
Information security policy
Incident response procedure
SOC 2 Type 1 report (from M9)
Annual pen test report (Enterprise)
Security architecture (deep dive)

Request a document Report a vulnerability