Frequently Asked Questions
Find answers to technical and operational questions about the Candlery event parsing and execution system.
Candlery is an AI-driven event intelligence engine built for Indian Equities. We monitor corporate regulatory feeds in real time, extract structured sentiment and directional intent from unstructured filings (like CAPEX expansion, profit guidance increases, or supply disruptions), and compile them into executable trading signals.
Candlery enforces a zero-hallucination policy using a deterministic 12-rule Risk Firewall. Any signal extracted by our cognitive LLM model pipeline must pass strict confidence thresholds (>= 95%), hold valid historical precedence analogs, and clear immediate liquidity and maximum exposure limits before compilation.
The LLM model acts as our core cognitive parsing layer. It extracts structured sentiment data from unstructured disclosures, processes intent parameters, and handles payload schema validation to prevent model drift and ensure zero-hallucination execution routing.
Yes. You can generate a sandbox API key directly on our home page. This key grants access to our simulated event streams and the interactive Sandbox Cockpit, allowing you to dry-run signals, test the Risk Firewall rules, and review latency telemetry.
Our pipeline continuously monitors official corporate filings from the BSE and NSE, annual and quarterly financial reports (10-K, 10-Q equivalent disclosures), insider trading activities, and corporate actions/board resolutions.