Trust Center

Built for enterprise trust.

An immutable GAAP ledger, cryptographic audit trail, and zero-trust architecture — not bolted on, but enforced at the database kernel level.

Evidence APIRequest SOC 2 Report
Live Security Statistics

Pulled from production database in real time — every number is verifiable via the Evidence API.

3
Immutable ledger events
tsm_ledger_events · tamper-proof triggers
0
State transitions audited
tsm_state_transitions · 10-year retention
183
Agent reasoning runs logged
proc_agent_runs · full reasoning captured
3
Approval decisions recorded
proc_approvals · named approver chain
Compliance Frameworks

Enterprise-grade evidence, not marketing copy.

Each framework is mapped to actual database controls, not checkbox compliance. Click any card to view the underlying evidence artifact.

🔐
SOC 2 Type II
CC-series controls mapped

Trust Services Criteria mapped to platform controls: CC6 (access), CC7 (change management), CC8 (risk), CC9 (vendor management). Immutable ledger satisfies evidence requirements.

View Security Posture
🛡️
ISO 27001
A-series annex controls

Annex A controls: A.9 (access control via RLS + service role), A.12 (operations — agent kill switch, health checks), A.14 (system acquisition — migration audit trail), A.16 (incident management — stuck event recovery).

View Compliance Mapping
🇪🇺
GDPR Article 30
Processing register

Records of processing activities: requestor PII (name, email, Telegram ID), supplier data (legal name, contacts, IBAN), invoice data. Data residency: Supabase EU region. Retention: 10 years for financial records per Dutch law.

View Processing Register
🔗
Cryptographic Hash Chain
SHA-256 linked audit trail

Every AI agent decision and treasury state transition is linked in a SHA-256 hash chain. Each row's hash encodes its predecessor — any tampering is instantly detectable via proc_check_chain_integrity(). Satisfies FDA 21 CFR Part 11 §11.70 (tamper-evident records) and SEC Rule 17a-4 (WORM audit trail). Most competitors claim immutability via policy; this proves it mathematically.

View Chain Integrity
⚔️
STRIDE Threat Model
Threat analysis & mitigations

Spoofing (JWT + service role segregation), Tampering (immutability triggers on ledger + state transitions), Repudiation (proc_agent_runs with full reasoning), Information Disclosure (RLS + service role), DoS (pg_net rate limiting), Elevation (role-based allowlists).

View Threat Model
Sub-Processors

Complete sub-processor register.

Sub-ProcessorRegionData CategoryLegal BasisCertification
SupabaseEU (Frankfurt)Database, Auth, Storage, Edge FunctionsContractual necessity (DPA available)SOC 2 Type II
OpenRouterUSLLM inference (agent reasoning only — no PII in prompts)Legitimate interest (pseudonymised payload)No PII transmitted
VercelUS (Edge CDN)Frontend hosting, serverless functionsContractual necessity (DPA available)SOC 2 Type II
Evidence API

Machine-readable compliance evidence.

All compliance evidence is available as a public API endpoint. Use it to feed your GRC tooling, SIEM, or audit dashboards.

Request
GET /api/trust/evidence?section=soc2&format=json
GET /api/trust/evidence?section=iso27001&format=csv
GET /api/trust/evidence?section=gdpr&format=json
GET /api/trust/evidence?section=all&format=json
GET /api/trust/evidence?section=soc2&from=2026-01-01&to=2026-03-31

# Rate limit: 10 requests/minute per IP
# No authentication required
Response
{
  "generated_at": "2026-04-11T10:00:00Z",
  "period": { "from": "2026-01-01", "to": "2026-03-31" },
  "framework": "soc2",
  "evidence": {
    "agent_success_rate_pct": 98.7,
    "approval_chain_records": 142,
    "ledger_event_count": 847,
    "last_state_transitions": [...]
  }
}
section=soc2section=iso27001section=gdprsection=all
Security Contact

Request our SOC 2 report or DPA.

Full SOC 2 Type II report and Data Processing Agreement (DPA) available to enterprise prospects under NDA. Security vulnerability disclosures welcome.

Security architecture in brief
Database kernel
All rules enforced at PostgreSQL level — no application bypass possible
Zero-trust agents
Whitelisted action model — agents cannot execute arbitrary SQL
Immutable ledger
BEFORE UPDATE/DELETE triggers raise exceptions — no record can be modified or deleted
Kill switch
Any agent can be disabled instantly — no code deployment required
Service role segregation
User sessions and service operations use separate JWT paths