AI & Agents
Quant Research
AI assistants query historical and live derivatives data conversationally — pulling IV surfaces, funding rates, open interest, and term structures on demand. Analysts ask questions in natural language and receive structured data responses grounded in real market data.
A derivatives analyst investigating a vol regime shift needs to pull ATM implied volatility across multiple tenors, check funding rates for directional bias, compare open interest by strike to gauge positioning, and cross-reference the term structure shape against historical precedents. Each data point typically requires a separate query, a different API endpoint, or a manual export from a dashboard.
The result is that the analytical cycle — from question to insight — takes longer than it should. By the time the data is assembled and the analysis is complete, the market may have already moved.
An AI research assistant connected via MCP can query derivatives data conversationally. "What’s the BTC 30-day ATM IV right now, and how does it compare to 7 days ago?" The assistant pulls the data, computes the comparison, and presents the answer — no endpoint mapping, no JSON parsing, no code.
The conversation retains context. "Now show me the same for ETH." "What does the term structure look like across 7d, 30d, 60d, 90d?" "Is the skew steeper than it was last month?" Each follow-up builds on the previous query. The assistant knows what asset, what tenor, and what timeframe the analyst is working with.
A single research question often spans multiple data types. "Is the current BTC vol regime consistent with what we saw before the March 2025 move?" Answering this requires historical IV surfaces, term structure snapshots, open interest distributions, and funding rate data — assembled from the same MCP tools in a single conversation flow.
Before entering a position, an analyst asks the assistant to pull the current vol surface, compare it to the 30-day average, identify which strikes are trading rich or cheap relative to the SVI-calibrated fair value, and flag any unusual skew. The assistant returns a structured assessment grounded in current data.
After closing a position, the analyst asks what happened to the vol surface between entry and exit. The assistant pulls the surface at both timestamps, computes the change in IV across relevant strikes, and attributes P&L to delta, gamma, vega, and theta movements using Greeks from calibrated surfaces.
Research teams producing weekly or daily market commentary can use an assistant to pull the latest data — vol surfaces, funding rates, open interest by strike, term structure shape — and generate structured commentary grounded in real numbers rather than generic observations.
The assistant can pull both the exchange-weighted composite surface and individual exchange-specific volatility surfaces from exchanges like Deribit, Bybit, OKX, GateIO, Derive and many more. An analyst investigating why Deribit vol is trading above the composite can query both in the same conversation, compare the divergence, and check whether it’s driven by a specific tenor or strike range.
Data available through the research assistant includes implied volatility surfaces (composite and per-exchange), options Greeks, funding rates, open interest, term structures, mark prices, index prices, and settlement prices. Coverage spans BTC, ETH, SOL, XRP, HYPE, ADA, SUI, and additional assets across Deribit, Bybit, OKX, GateIO, and additional venues. Historical and real-time data are accessible through the same conversational interface.
