API-first architecture, developer-friendly integration, and no-code rule management — so your risk team can iterate in hours, not sprint cycles.
Single POST to /transactions/score returns a fraud decision, score, SHAP explanations, and triggered rules in one response. No batch jobs, no polling. Integrates into any payment flow in hours — not weeks. Full OpenAPI documentation included.
vcurd's swarm agent continuously analyzes flagged transactions and suggests new rule conditions based on emerging patterns. Your risk team reviews and activates suggestions without writing code — turning fraud signal into detection coverage in minutes.
Before disbursal, score the full applicant pool in batch mode. vcurd's batch API processes CSV uploads with async job tracking — returning mule scores, bureau velocity checks, and ML fraud scores for every applicant before funds leave your system.
Switch between OpenAI GPT-4o, Claude, Gemini, and local Ollama models from the admin panel — no code changes. Use different models for different risk tiers: fast local model for real-time scoring, GPT-4o for deep fraud investigation summaries.
The scoring API returns everything you need — decision, score, triggered rules, and SHAP explanations — in a single synchronous response.
API keys and documentation available immediately after onboarding.