Overview

The Block Scholes MCP (Model Context Protocol) server lets you query crypto derivatives data and run strategy backtests directly from AI assistants. Instead of writing API calls or building custom infrastructure, you describe what you need in natural language and the MCP handles data retrieval, instrument discovery, and trade backtesting behind the scenes in a safe, deterministic manner.

Why Deterministic?

AI assistants are powerful reasoning tools, but left to their own devices they can hallucinate data — fabricating prices, implied volatilities, or backtest results that look plausible but are entirely wrong. The MCP solves this by grounding every response in real data from the Block Scholes platform. When your assistant needs a number, it calls a tool and gets a verified result — it doesn't guess.

This means identical queries always produce identical outputs. Whether you're pulling an IV surface today or re-running the same query next month for audit purposes, the numbers won't drift. For financial modelling, risk management, and compliance workflows, that reproducibility isn't optional it's the whole point.

How it Works

The Model Context Protocol (MCP) is an open standard designed to connect AI assistants with external data sources. Your AI client — whether Claude, Cursor, or OpenClaw — establishes an authenticated session with the Block Scholes MCP server. From that point on, the assistant can call Block Scholes tools on your behalf: retrieving market data, looking up instruments, or executing backtests, all through natural language.

The server sits between your AI assistant and the Block Scholes data platform, translating conversational requests into structured queries. This means you get the full depth of our derivatives data including implied volatility surfaces, forward curves, funding rates, open interest, and more. The following tools are defined

Tool	Description
data_retrieval	Query spot, perpetual, futures, and options price data. Retrieve implied volatility surfaces, mark prices, funding rates, open interest, volumes, and more across multiple exchanges.
backtest_strategy	Execute strategy backtests with trade-by-trade PnL, hedge attribution, and funding rate incorporation. Supports spot, perpetual, futures, and options positions with theoretical or listed pricing modes. Requires additional permissions or Strategy Backtest subscription.
get_instruments	Look up available instruments per exchange, including listed expiries, strikes, and option types.

Access Comprehensive Datasets with Block Scholes

Unlock institutional-grade datasets with Block Scholes, providing an extensive range of data across spot, perpetual, futures, and options markets. These datasets are sourced from leading exchanges such as Deribit, OKX, and Bybit, ensuring high-quality data for informed decision-making. Our offerings are perfect for thorough analysis on a variety of trading instruments.

Data Coverage

• Spot Markets: Gain insights into real-time trading activities with spot market data.

• Perpetual Markets: Stay ahead with perpetual contract data that tracks ongoing trading trends.

• Futures Markets: Understand future price directions with detailed futures data analytics.

• Options Markets: Enhance your strategies with comprehensive options data.

• RWA(*): FX, Commodities, Equities and Indices (Upon request)

Subscription Information

To explore or customise a subscription suited to your needs, please visit our Data Plans. Our options are designed to support various research and trading needs, ensuring optimal benefit from your data-driven activities.

To get started with our Backtesting MCP package, visit here.

Backtest Strategy Tool Overview

The backtest_strategy tool supports multi-leg positions across various asset types. Key features include:

• Ability to backtest spot, perpetual, futures, and options trades in a single call.

• Use of theoretical (model-derived) or listed (market-implied) pricing.

• Inclusion of delta hedging via spot or perpetual instruments.

• Customisation of entry/exit dates, tenors, strikes, and quantities.

Results provide detailed trade-by-trade P&L with comprehensive hedge attribution, alongside trade analytics such as drawdown and Sortino ratio, making the results ideal for tabular or interactive outputs.

Supported Clients

Block Scholes MCP works with any client that supports the MCP protocol. We provide setup guides for:

• Claude (Web and Desktop) — Recommended

• Cursor

• Other MCP Clients

Authentication

All MCP connections authenticate via OAuth. When you connect for the first time, you will be prompted to sign in or create a Block Scholes account. Your data access is determined by your subscription plan.

Requirements

• A Block Scholes account with an active subscription

• An MCP-compatible AI client (see supported clients above)

Add the Block Scholes skill

After connecting a client, create the shared Block Scholes Skill.

This keeps crypto derivatives requests, backtests, and output formatting consistent across supported clients.

For best results, use Claude Opus 4.5 or above with code execution enabled. This ensures accurate tool-calling and richer data visualisation.

Next Steps

Head to Connecting to set up your preferred client, then create the shared Block Scholes Skill. For dataset coverage and query details, see Data Access.