MARKET SENTRY

Inspiration

The financial world moves at lightning speed, but retail investors are often the last to know the truth. Every year billions are lost to coordinated pump and dump schemes that artificially inflate asset prices only to crash them later. While traditional analytics and alerting tools exist, they only surface trends; they do not understand context or intent. Our team witnessed friends and students losing money due to misleading hype, and we asked a simple question: What if investors had an autonomous AI analyst that could tell them whether the hype behind a stock or crypto asset was real or manipulated in real time?

What it does

MARKET SENTRY is an autonomous AI market analyst that identifies manipulation behind sudden price surges and social media hype. It moves beyond basic alerting systems by researching, reasoning, and delivering a clear verdict: JUSTIFIED or UNJUSTIFIED. It analyzes fast-moving social activity, gathers ground-truth facts using RAG from official and credible sources, and compares hype signals with factual data. If manipulation is detected, it issues a red alert within seconds, allowing exchanges or regulators to act before investors are harmed.

How we built it

We designed a three-agent AI workflow that behaves like a human research and analysis team. The Watchdog Agent continuously monitors social media to detect rapid coordinated spikes. When unusual patterns are detected, the Research Agent is activated and performs Retrieval-Augmented Generation to collect official news, filings, whitepapers, and press releases. Finally, the Analyst Agent compares the hype against reality using generative reasoning and produces a final verdict. The architecture follows a two-speed model where the lightweight watchdog processes 99.9 percent of the incoming data cheaply, and the expensive analyst model runs only on confirmed triggers, making the system scalable and cost-efficient.

Challenges we ran into

Collecting live data at speed and filtering real manipulation signals from noise was technically demanding. Designing clean triggers that avoid false positives required significant experimentation. Integrating RAG for financial data sources while maintaining latency constraints was complex. Building a reasoning layer capable of generating consistent verdicts and measurable confidence levels was another major challenge. Balancing accuracy and performance was a constant engineering decision.

Accomplishments that we're proud of

We successfully built a working end-to-end pipeline that autonomously detects speculative manipulation and produces verdicts grounded in real data. The three-agent architecture performed reliably during stress scenarios, and our initial simulations demonstrated that MARKET SENTRY can flag fraudulent activity significantly faster than existing monitoring tools. We are proud that our solution provides real protection for real investors, something that can have a meaningful societal and financial impact.

What we learned

We learned the importance of designing AI systems that think and analyze, not just predict. We learned how powerful autonomous multi-agent systems can be when orchestrated correctly. Implementing RAG in a real-time decision-making environment taught us how to balance cost, latency, and reasoning depth. We also realized that financial AI must be transparent and explainable to build trust with institutions and regulators.

What's next for MARKET SENTRY

We plan to integrate live exchange APIs and build a real-time B2B dashboard for financial regulators and trading platforms such as SEBI, Zerodha, and Binance. We aim to support multilingual global markets and expand coverage beyond crypto and equity assets. Our long-term vision is to launch MARKET SENTRY as a SaaS risk-intelligence API that becomes the industry standard for fraud detection and hype validation at scale.