Corint combines Corint Decision, a real-time risk decision engine that executes RDL strategies, with Corint Agent, an AI agent for risk analysis, model optimization, rule strategy generation, backtesting, and deployment.
Open-source core today. Managed cloud and team workflows as you scale.
Two products, one operating loop
Teams can adopt Corint Decision as infrastructure, Corint Agent as the operating layer, or both together as a closed-loop risk platform.
Real-time decision engine
Deterministic risk execution for production systems.
Define rules, features, rulesets, and pipelines in RDL, then serve millisecond-level decisions through a high-performance Rust engine with audit trails and observability built in.
AI agent for risk operations
Analyze risk, improve models, and optimize strategies with AI.
Investigate anomalies, diagnose emerging risk patterns, build or tune risk models, and generate or optimize rule strategies through natural language workflows.
Corint Decision
Corint Decision is not a generic rule framework. It is a deployable decision system with a unified RDL, external APIs, feature evaluation, observability, and explainable decisions designed for financial-grade risk workflows.
<10ms
p99 target powered by a Rust decision runtime
RDL
One language for rules, features, and pipelines
Evaluate rules for online events with deterministic latency and production-grade concurrency.
Run as open-source infrastructure or consume it as managed Corint Cloud services.
risk_pipeline.rdl.yaml
Unified DSLrule:
id: high_velocity
when:
all:
- transaction_count_1h > 10
score: 75
ruleset:
id: fraud_detection
rules: [high_velocity]
conclusion:
- when: total_score >= 100
signal: decline
pipeline:
id: payment_risk
steps:
- step:
type: ruleset
ruleset: fraud_detectionCorint Agent
Corint Agent helps risk teams investigate anomalies, identify risk drivers, build or tune models, generate rule strategies, and optimize existing policies with backtesting, simulation, approvals, and auditability built into the workflow.
Core workflow
Explain spikes in reject rate, abnormal transaction clusters, feature drift, risky user segments, and emerging fraud patterns.
Assist with feature discovery, model performance diagnosis, threshold tuning, score-band analysis, and model comparison.
Turn risk hypotheses into RDL rules, rulesets, and pipelines that can be reviewed, tested, and deployed.
Find redundant rules, tune thresholds, reduce false positives, and balance approval rate, risk cost, and manual review load.
Replay historical decisions, compare before/after impact, and validate a change before it reaches production traffic.
Produce role-aware explanations, audit-ready reasoning, monitoring summaries, and recurring risk operation reports.
How they work together
Describe a question, incident, or strategy goal in natural language.
Analyze anomalies, tune models, generate RDL strategies, and validate impact.
Execute the approved RDL in production with deterministic performance.
Use traces, outcomes, and experiments to improve the next strategy.
Use cases
Credit approval, limit management, transaction monitoring, identity risk, and default prevention.
Card testing, abnormal velocity, account takeover, AML signals, and cross-border payment risk.
Promotion abuse, fake transactions, chargeback risk, seller fraud, and return abuse.
API abuse, account sharing, trial abuse, billing fraud, and resource abuse detection.
Open source first
The decision engine is the open-source foundation: deploy it locally, inspect the RDL model, wire it into your systems, and build confidence before adopting managed workflows.
Explore Corint DecisionManaged cloud path
Managed Decision APIs, Agent workflows, team collaboration, monitoring, SLA guarantees, and integrations for data warehouses, feature stores, and approval systems.
Join Cloud WaitlistInsights
Product essays on risk intelligence, RDL, and the shift from rule maintenance to agent-assisted risk operations.
From rule-based systems to AI-powered cognitive intelligence.
Read more ->
The technical philosophy behind one language for features, rules, and decisions.
Read more ->
How agent-assisted strategy generation changes risk operations.
Read more ->Adopt Corint Decision as open-source risk infrastructure, evaluate Corint Agent for AI-assisted operations, or use both as a single loop for strategy development and real-time execution.