ProductUse CasesDevelopersCompanyBlogRequest Access

Log your first AI decision in 5 minutes.

A single REST API endpoint. SDKs for Python, Node.js, and Java. Comprehensive documentation. Built for engineering teams integrating AI traceability into production systems.

One integration. Complete decision provenance.

PythonNode.jsJava
# Install the SDK
pip install aegis-trace

# Log a decision
from aegis_trace import AegisClient

client = AegisClient(api_key="at_live_...")
record = client.log_decision(
    agent_id="suitability-engine-v3",
    decision_type="portfolio_recommendation",
    inputs={
        "client_risk_profile": "moderate",
        "market_conditions": "volatile"
    },
    output={
        "action": "REDUCE_EXPOSURE",
        "fund": "GB00B3X7QG63"
    },
    regulatory_context=["FCA_CONSUMER_DUTY", "MIFID_II"]
)

print(record.certificate_id)  # AT-2026-07-14-c3a9f1e
print(record.verdict)          # PASS
print(record.hmac_signature)   # b7e2d4c8...f19a03e1

How the pipeline works.

01
Decision payload sent via REST API or SDK
02
PII redaction via Presidio sidecar (34 recognisers, 13 EU countries, 9 languages), runs inside client network
03
HMAC-SHA256 signing with managed KMS key infrastructure
04
Tamper-proof storage with configurable retention policy (7-year default)
05
Certificate retrieval via API, sub-200ms p99 latency
Infrastructure
Compute
Google Kubernetes Engine (GKE), autoscaling, multi-zone availability
Database
AlloyDB (PostgreSQL-compatible), managed backups, point-in-time recovery
Key Management
Google Cloud KMS, hardware security modules, automatic key rotation
Data Residency
europe-west2 (London), EU data processing, GDPR compliant
Security
mTLS for internal service communication, RFC 3161 timestamping
Latency
Sub-200ms p99 certificate retrieval, sub-2s end-to-end sealing

Key endpoints.

POST/v1/decisionsLog a decision
GET/v1/decisions/{id}Retrieve a decision record
GET/v1/decisionsList decisions (paginated, filterable)
GET/v1/export/fcaExport FCA Consumer Duty format
GET/v1/export/euaiExport EU AI Act Article 12 format
POST/v1/verify/{id}Verify certificate integrity

Deployment options.

Cloud API
Managed Service
REST endpoint with OpenAPI 3.0 spec. Rate limits: 1,000 req/min (standard), custom limits available. 99.9% SLA. Auto-scaling. No infrastructure to manage.
Private Cloud
Containerised
Docker/Kubernetes deployment. Helm charts provided. Environment variables for configuration. Managed updates via container registry. AWS, GCP, or Azure.
On-Premises
Air-Gapped
Full deployment package. Offline installation supported. Network requirements: outbound HTTPS for updates (optional). Minimum: 4 vCPU, 16GB RAM, 500GB storage.

Client libraries and integration.

Python SDK
Available via PyPI. Python 3.8+. Async support. Type hints throughout.
pip install aegis-trace
Node.js SDK
Available via npm. Node 18+. TypeScript definitions included. ESM and CJS.
npm install @aegis-trace/sdk
Java SDK
Available via Maven Central. Java 17+. Spring Boot integration module available.
com.aegistrace:aegis-sdk
REST API
OpenAPI 3.0 specification. Language-agnostic. Full API reference provided on access.
Webhook Support
Configure webhooks for FLAG and FAIL verdicts. Real-time notifications to your compliance or incident management systems.

Request API access and technical documentation.