13 Modules
Platform Overview
Every module in the Trustinera AI platform — tap any card to jump to its full reference.
Platform
Core infrastructure, authentication, RBAC, audit trails, alerts, signals, guard rails, OpsIQ operational intelligence, infrastructure stores, orchestration, and the AI Chat subsystem.
Capabilities
- Identity & access management with fine-grained RBAC
- Immutable audit trail for every platform action
- Signal and signpost framework for human-in-the-loop decisions
- Guard rails engine enforcing policy gates on all AI outputs
- Alert pipeline with escalation policies and notification channels
- Infrastructure stores: notebooks, vector DB, object storage, data endpoints
- Helper orchestration layer (LangGraph-based agent runtime)
- Distributed tracing across all microservices
- Kong API Gateway with 800+ route definitions
- Theme and sidebar customisation per organisation
Wizards
Alerts (5)
Monitoring (6)
Automation (5)
Verification (14)
Evaluation (14)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/auth/* |
Authentication & token management |
/api/users/*, /api/organizations/*, /api/roles |
Identity & RBAC |
/api/alerts/*, /api/alert-history |
Alert pipeline |
/api/signals/*, /api/signposts/* |
Signals & HITL signposts |
/api/guard-rails/* |
Guard rails engine |
/api/audit/*, /api/logs/* |
Audit & logging |
/api/store/* |
Infrastructure stores (notebooks, vector, bucket, endpoints, ingestion) |
/api/orchestration/* |
Helper orchestration |
/api/platform/* |
Platform configuration, RBAC, helpers |
/api/chat/* |
AI Chat (via AI Gateway) |
/api/trace/* |
Distributed tracing |
Categorise
AI-powered transaction classification with a 3-tier ML pipeline (rules, classical ML, LLM), HMRC compliance mapping, multi-currency support, and continuous learning from human corrections.
Capabilities
- 3-tier classification pipeline: rules, Naive Bayes / gradient boosted models, LLM fallback
- 95-99% accuracy on UK financial transaction data out of the box
- HMRC VAT category and tax code mapping
- Confidence scoring with configurable human-review thresholds
- Continuous learning from corrections via feedback loop
- Batch and real-time categorisation modes
- Split transaction and recurring pattern detection
- A/B testing framework for model experiments
- Keyword-based and merchant-based rule engines
- Full explainability on every categorisation decision
Wizards
Creation (10)
Training (5)
Verification (10)
Evaluation (15)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/transactions/* |
Transaction CRUD and categorisation |
/api/categories/* |
Category management |
/api/upload |
File ingestion (CSV, OFX, QIF) |
/api/feedback |
Correction / feedback loop |
/api/training |
Model training pipeline |
/api/bank |
Bank feed management |
/api/dashboard-stats |
Categorise analytics |
/api/accuracy-history |
Accuracy tracking |
Sentrise
Real-time risk intelligence and fraud detection with 8-factor risk scoring, sanctions screening, entity network analysis, AML/KYC compliance, and velocity-based fraud rules.
Capabilities
- 8-factor composite risk scoring (amount, velocity, geography, entity, device, pattern, network, time)
- Sub-100ms sanctions screening against OFAC, HM Treasury, UN, EU, and custom watchlists
- Entity network analysis for detecting fraud rings
- Velocity-based fraud rules with configurable time windows
- PEP (Politically Exposed Person) screening
- Device fingerprint and behavioural anomaly detection
- SAR-ready case file generation
- Real-time and batch risk scoring modes
- Investigation case management with evidence chains
- Block rules with instant enforcement
Wizards
Creation (15)
Verification (10)
Evaluation (15)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/sentrise/* |
Risk intelligence API surface |
/api/live/* |
Live risk status and streaming |
TrustFlow
Complete ML/AI lifecycle management with a 14-phase pipeline, 7 learning paradigms, Helper architecture for composing AI agents, artifact versioning, experiment tracking, and policy-gated promotion.
Capabilities
- 14-phase ML pipeline: Ingest, Transform, Map, Enrich, Feature, Train, Evaluate, Experiment, Deploy, Monitor, Automate, Comply, Operate, Observe
- 7 learning paradigms: supervised, unsupervised, semi-supervised, reinforcement, transfer, few-shot, self-supervised
- Helper architecture for composing multi-component AI agents
- Artifact lifecycle: Draft, Deployed, Released
- Experiment tracking with hyperparameter sweeps
- Model bias audit and fairness checks
- Prompt template management for LLM-based components
- Notebook integration (Jupyter) for exploratory work
- Canary release with traffic splitting
- Full model lineage and reproducibility
Wizards
Data & Features (12)
Training & Evaluation (7)
Deployment (9)
Operations (11)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/trustflow/* |
Full TrustFlow API surface (artifacts, pipelines, helpers, experiments) |
Reconcile
Financial operations module covering bank reconciliation, invoicing, expense management, P&L reporting, CRM, and AI-powered matching that reduces period close from 5 days to 1.
Capabilities
- AI-powered bank reconciliation with fuzzy matching
- Invoice creation, recurring invoices, credit notes, and quotes
- Expense claims and batch payment processing
- Chart of accounts and opening balance management
- VAT return preparation and verification
- Aged debtors and aged creditors reporting
- Profit & loss and balance sheet generation
- Intercompany balance reconciliation
- Bank feed connection management
- Duplicate entry detection and prevention
Wizards
Banking (5)
Invoicing (5)
Payments (5)
Verification (10)
Evaluation (15)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/reconcile/* |
Reconciliation, invoicing, payments, reporting |
Conform
Compliance management across SOC 2, ISO 27001, GDPR, EU AI Act, and custom frameworks with automated evidence collection, control testing, audit scheduling, and regulatory change tracking.
Capabilities
- Multi-framework compliance: SOC 2, ISO 27001, GDPR, EU AI Act, PCI DSS, DORA
- Automated control testing and effectiveness scoring
- Evidence collection and management with chain-of-custody
- Audit scheduling with finding tracking and remediation plans
- Policy lifecycle management with version control
- Vendor risk assessment and compliance monitoring
- Compliance incident management and response
- Training tracking and completion verification
- Cross-framework control mapping (one control, many frameworks)
- Board-ready compliance reporting and maturity scoring
Wizards
Framework (5)
Risk & Policy (5)
Audit (5)
Verification (10)
Evaluation (15)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/conform/* |
Compliance frameworks, controls, audits, policies, evidence |
Contain
Infrastructure management with server provisioning, Kubernetes cluster orchestration, network configuration, 90+ security hardening tools, and one-click Kubernetes deployment.
Capabilities
- Server provisioning across bare-metal, VM, and cloud providers
- Kubernetes cluster creation and node pool management
- Network and subnet management with firewall rules
- SSL certificate management and DNS resolution verification
- Auto-scaling policies with configurable triggers
- Storage volume provisioning and snapshot management
- Load balancer configuration and health checks
- 90+ security hardening checks (CIS benchmarks)
- Backup integrity verification and disaster recovery planning
- Container image scanning and vulnerability assessment
Wizards
Provisioning (10)
Networking (5)
Verification (10)
Evaluation (15)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/contain/* |
Servers, clusters, nodes, networks, storage, security |
LogiFlow
Data pipeline orchestration with ETL, streaming, data contracts, schema validation, quality gates, DAG-based execution, quarantine management, and SLA-driven monitoring.
Capabilities
- Visual flow designer for ETL/ELT pipelines
- Streaming and batch data processing modes
- Data contracts with schema drift detection
- Quality gates with configurable thresholds
- DAG-based execution with parallel processing
- Quarantine management for failed or suspect records
- Schema definition and validation with lineage tracking
- SLA policies with automatic alerts on breach
- Retry policies with exponential backoff
- Approval gates for production data changes
Wizards
Pipeline (10)
Transforms (10)
Verification (10)
Evaluation (11)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/logiflow/* |
Flows, transforms, contracts, schemas, schedules, quarantine |
Observe
Business intelligence dashboards with 14 widget types, scheduled reports, data connections (SQL and API), saved queries, executive summaries, and real-time data feeds.
Capabilities
- 14 widget types: line, bar, pie, area, gauge, table, KPI, heatmap, scatter, funnel, map, text, metric, real-time feed
- Drag-and-drop dashboard builder
- SQL and API data connections
- Saved and scheduled query execution
- Scheduled report generation and distribution
- Executive summary auto-generation
- PDF and CSV export for all dashboards
- Dashboard sharing with role-based access
- Template library for quick dashboard creation
- Anomaly alert widgets with configurable thresholds
Wizards
Dashboards (5)
Connections (4)
Queries (4)
Reports (4)
Advanced (5)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/observe/* |
Dashboards, widgets, queries, connections, reports |
Cost Intelligence
Cost tracking, budgeting, forecasting, and optimisation across all platform resources with chargeback/showback, unit economics, anomaly detection, and sustainability scoring.
Capabilities
- Cost centre management with hierarchical rollups
- Resource tagging and tag policy enforcement
- Rate card management for internal pricing
- Budget creation with spending alerts
- Forecast rules with multiple model types
- Chargeback and showback reporting
- Unit economics calculation per module
- Cost anomaly detection with root cause analysis
- Idle resource identification and optimisation
- Reserved instance and right-sizing recommendations
Wizards
Setup (5)
Budgets (10)
Verification (10)
Optimisation (6)
Evaluation (9)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/cost-intelligence/* |
Cost centres, budgets, forecasts, chargebacks, optimisation |
Insight Mesh
Cross-module intelligence layer providing dependency mapping, signal routing, correlation analysis, SLA chains, blast radius modelling, and unified topology views across the entire platform.
Capabilities
- Cross-module data correlation and pattern detection
- Dependency mapping with blast radius analysis
- SLA chain monitoring with breach forecasting
- Event bridge for cross-module event propagation
- Resource group management for logical clustering
- Topology views with real-time health aggregation
- Cascade failure simulation
- Signal route management for inter-module communication
- Natural language mesh querying
- Cross-module data lineage tracking
Wizards
Mesh (10)
Topology (5)
Verification (10)
Analysis (15)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/insight-mesh/* |
Mesh links, signal routes, correlations, topology, SLA chains |
Trust Score
Unified quality scoring across data, models, compliance, and operations with drift monitors, fairness checks, model scorecards, stakeholder trust reports, and automated remediation.
Capabilities
- Composite TrustScore across data quality, model quality, compliance, and security
- Quality rule engine with configurable thresholds
- Quality profiles for different data domains
- Quality baselines with drift detection
- Model scorecards with multi-dimensional grading
- Drift monitors for data and model distribution shifts
- Fairness checks across protected characteristics
- Explainability configuration and verification
- GDPR assessment automation
- Stakeholder trust reporting and trend analysis
Wizards
Quality Rules (5)
Model Quality (5)
Compliance (5)
Verification (10)
Analysis (15)
AI-Powered (4)
Helpers / Tools
API Endpoints
| Route | Description |
|---|---|
/api/trust-score/* |
Quality rules, scorecards, drift monitors, fairness, compliance |
AI Chat
Platform-owned conversational AI interface providing natural language access to every Helper, MCP tool discovery, contextual awareness across modules, and full audit trail on every interaction.
Capabilities
- Natural language interaction with all platform Helpers
- MCP (Model Context Protocol) tool discovery and execution
- Dual-mode Helper execution: structured JSON and free-form chat
- Context-aware conversations with module state awareness
- Chat history persistence in tai_chatbot schema
- Guard rails applied to all chat interactions
- Explainability on every AI response
- Signpost generation for low-confidence or high-impact responses
- Multi-turn conversations with memory
- Audit trail on every chat interaction
API Endpoints
| Route | Description |
|---|---|
/api/chat/* |
Conversational AI (routed via AI Gateway to LLM) |
Helper Store
48 production-ready AI Helpers.
Every Helper supports dual-mode execution — structured JSON for programmatic use and free-form chat for conversational interaction. All include explainability, audit trails, and HITL signposts.
Risk & Fraud (8)
8-factor composite risk score on any transaction or entity in real time.
Identifies emerging fraud patterns using graph analysis and sequence modelling.
Sub-100ms screening against OFAC, HM Treasury, UN, EU, and custom watchlists.
Continuous anti-money laundering monitoring with threshold-based alerting.
Detects velocity anomalies across configurable time windows and dimensions.
Analyses identity documents and video for deepfake artefacts.
Multi-factor identity verification combining document, biometric, and data checks.
Profiles user behaviour and flags deviations from established patterns.
ML & AI (8)
UK tax-compliant transaction categorisation with HMRC code mapping.
3-tier ML pipeline (rules, classical ML, LLM) with 95-99% accuracy.
One-click Naive Bayes model training on labelled transaction data.
GPU-accelerated deep learning training for complex classification tasks.
Generates vector embeddings for text, transactions, and entities.
Continuous monitoring of prediction distributions against training baselines.
Automated feature extraction and transformation from raw data.
Bayesian and grid search optimisation of model hyperparameters.
Compliance (8)
Automated risk classification and compliance gap analysis for EU AI Act.
Automated evidence gathering for SOC 2 Type II audit controls.
Continuous monitoring of personal data processing against GDPR requirements.
Versioned policy management with approval workflows and distribution tracking.
Pattern analysis across audit logs to detect anomalous access or behaviour.
Quantitative scoring of control effectiveness based on test results and incidents.
Third-party risk assessment with questionnaire automation and scoring.
Monitors regulatory publications and maps changes to affected controls.
Data & Integration (8)
Robust CSV parsing with encoding detection, type inference, and quality checks.
Parses OFX, QIF, CSV, and MT940 bank statement formats.
Bi-directional synchronisation with Xero accounting platform.
Configurable API polling with pagination, auth, and retry handling.
Validates data against JSON Schema, Avro, or Protobuf definitions.
Multi-dimension data quality scoring: completeness, accuracy, consistency, timeliness.
Configurable transform engine with 50+ built-in transform functions.
Fuzzy and exact matching deduplication with configurable merge strategies.
Operations (8)
Continuous health checking across all platform services with dependency mapping.
Groups related alerts into incidents using temporal and topological correlation.
Pattern extraction and anomaly detection across structured and unstructured logs.
Identifies idle resources, right-sizing opportunities, and reserved instance savings.
Tracks SLA compliance across all services with breach prediction.
Classifies incidents by type, severity, and affected service using ML.
Forecasts resource requirements based on growth trends and seasonal patterns.
Profiles API latency, throughput, and error rates across all endpoints.
Intelligence (8)
Graph-based analysis of entity relationships for fraud ring detection.
Identifies correlations between signals across different modules.
Time-series forecasting for any metric using ensemble models.
Traces anomalies back through dependency chains to identify root causes.
Generates human-readable reports from structured data and metrics.
Creates optimised dashboards from natural language descriptions of KPIs.
Content and action recommendations based on user behaviour and context.
Automatically constructs knowledge graphs from unstructured data sources.
At a Glance
Platform Capabilities Summary
Key Differentiators
Every AI decision includes a human-readable explanation of the reasoning chain, confidence drivers, and data evidence. Not optional — mandatory on every Helper execution.
Low-confidence decisions, guard rail violations, high-value actions, and first-time patterns automatically generate signposts for human review. HITL gates block execution until approved.
Every execution across every module is logged to OpsIQ with full input/output, component execution order, guard rail results, and distributed trace IDs. FCA-ready from day one.
Every Helper supports both structured JSON tool execution and free-form natural language chat — the same pattern as MCP Tools. Programmatic and conversational from a single endpoint.
Per-module database roles, schema-level access control, Kong API Gateway routing, and the Destination Abstraction Layer ensure no module can access another module's data.
HMRC tax code mapping, FCA Consumer Duty compliance, PSD2 open banking, BACS/CHAPS/Faster Payments support, and UK data residency — purpose-built for regulated UK financial environments.
Go deeper