> ## Documentation Index
> Fetch the complete documentation index at: https://docs.backquant.com/llms.txt
> Use this file to discover all available pages before exploring further.

# What to ask it

> Example questions for the BackQuant MCP server, and which tool each one reaches for.

The server is designed to be used in plain language. You do not name tools; your
AI client picks them. These are questions that work well, grouped by what you
are trying to find out.

<Note>
  Everything here is **crypto**. BTC and ETH have the deepest coverage; SOL and
  HYPE are supported where data exists.
</Note>

## Start here

Good opening questions when you want a read on the market:

> **"What are the gamma levels on BTC right now?"**
> Returns the gamma flip, call wall and put support, with spot for context.
> *(`get_gex_levels`)*

> **"Compare BTC and ETH dealer positioning."**
> Both markets side by side in one call, rather than asking twice.
> *(`compare_symbols`)*

> **"Which strikes carry the most gamma on ETH?"**
> The strike-by-strike breakdown behind the headline levels.
> *(`get_gex_strike_profile`)*

## What is actually trading

> **"What has been trading on ETH today?"**
> Recent prints and aggregates over them.
> *(`get_tape`)*

> **"Where did aggressors buy and sell calls on BTC this week?"**
> Bought against sold, strike by strike, on still-open contracts.
> *(`get_flow_by_strike`)*

> **"Show me the biggest BTC options trades today."**
> Whale prints, filtered by premium.
> *(`get_tape`)*

> **"Has dealer hedging pressure been building on BTC?"**
> Hedging pressure over time rather than a single snapshot.
> *(`get_dealer_flow`)*

## Volatility and what the market is pricing

> **"What move is the market pricing for the next BTC expiry?"**
> The implied move from ATM volatility for that expiry.
> *(`get_expected_move`)*

> **"What are the odds BTC finishes above 70k this month?"**
> Market-implied probability of finishing above or below a level.
> *(`get_probability`)*

> **"How does ETH implied vol look across expiries?"**
> The term structure, plus smile and skew cuts if you ask for them.
> *(`get_iv`)*

## Structure and expiries

> **"What expiries are listed for BTC?"**
> Valid expiry tokens, useful before asking anything expiry-specific.
> *(`list_expiries`)*

> **"How much open interest rolls off at the next BTC expiry?"**
> Upcoming expiries and the size expiring at each.
> *(`get_opex`)*

> **"Where is max pain for the front ETH expiry?"**
> *(`get_max_pain`)*

> **"Give me an overview of every BTC expiry."**
> Dates, days to expiry, ATM vol and open interest in one table.
> *(`get_expiry_summary`)*

## Over time

> **"How has BTC gamma changed over the last two weeks?"**
> A metric's history rather than its current value.
> *(`get_history`)*

> **"Has open interest been rising on ETH?"**
> *(`get_oi`)*

## Getting better answers

**Name the market.** "BTC" or "ETH" beats "bitcoin's options" for getting the
right tool called first time.

**Say the timeframe out loud.** "over the last two weeks" produces a history
call; without it you get a snapshot of right now.

**Ask for one thing at a time.** A question bundling levels, flow and vol
becomes three tool calls, each drawing on your quota. Ask, read, follow up.

**Use real expiry tokens.** If you want a specific expiry, ask for the list
first, then reference one. The model can also do this itself in two steps.

**Ask it to explain its answer.** The tools return numbers; the interpretation
is your client's. "Why does that matter for spot?" usually produces a better
read than asking for the interpretation up front.

## What it will not do

* **Equities.** The API is crypto-only, so no SPX, NDX or single stocks.
* **Trade for you.** It is read-only market data. There are no order tools.
* **See your screen.** Unlike the AI panel in the terminal, MCP has no view of
  your workspace or panels. It only knows what it queries.
* **Predict.** It reports positioning, flow and pricing. Any forecast is your
  client reasoning over those numbers, not a BackQuant output.