Watch
ai assistant
Coming soon · Beta

option.watch AI — your options copilot

Scans live markets. Watches your positions. Teaches you as you trade.

The bicycle problem

Learning to ride a bicycle is hard at first — not because any single thing is difficult, but because you must control balance, steering, pedalling, and braking simultaneously, before any of them are automatic. Get one wrong and you fall.

Options are exactly the same. Strike selection, expiry, IV level, delta, theta decay, position sizing — each one is learnable in isolation. Together, under real market pressure, with real money, they overwhelm. Most new options traders blow up not from a single mistake but from losing track of the picture while looking at one variable.

A good cycling trainer runs beside you, steadies the seat, and talks you through what to feel for. option.watch AI is that trainer for options. It is always watching, always knows the full picture, and speaks up when something needs your attention.

What the assistant sees

Unlike a generic chatbot that gives textbook answers, this AI has live market data embedded in every response — not from a data vendor API, but directly from the same order book engine that powers the charts.

Live market data

Current bid/ask at every strike, IV curve, skew, underlying price — all exchanges simultaneously, updated every 100ms.

Options flow

Recent large trades across all instruments. When a big player prints a block, the assistant knows — and can tell you what chain they traded in.

Your positions

If a trading account is connected, the assistant sees your open positions, entry prices, current P&L, delta exposure, and days to expiry.

IV and skew

Implied volatility at the money, the put-call skew across strikes, and the term structure across expiries — the full picture of what the market is pricing.

Prediction market bridge

Live Polymarket contracts alongside option chains. The AI translates plain-English prediction market questions into regulated option trades — and compares implied probabilities between the two.

What it can do

1 — Answer any options question

Ask anything from first principles to advanced strategy. The assistant knows options theory deeply, but always answers in the context of what is live on your screen.

conversation example
what is IV skew and why does it matter?
IV skew is the pattern of implied volatility across different strikes. A typical equity skew has OTM puts trading at higher IV than OTM calls — because investors pay more for downside protection than upside participation.

Right now on SPY: the May 560 put has IV ~23%, the May 600 call has IV ~16%. That 7-point gap is the skew. If you sell the 560 put and buy the 600 call (a risk reversal), you collect the difference — but you're short a downside move. Want me to show you the current strikes?

2 — Scan the market: "what should I trade today?"

The assistant scans every chain across all exchanges — crypto and US equities — and looks for elevated IV, unusual skew, abnormal flow, and cross-exchange mispricings. It reasons about the full picture and proposes specific, actionable ideas.

conversation example
what should I trade today?
scan_market() — reading all chains across 8 exchanges
Three things stand out this morning:

BTC Deribit Jun27: ATM IV at 84%, put skew +13vol at 90% moneyness. Someone is paying heavily for downside. If you expect BTC to stay range-bound, the risk reversal (sell 90k put / buy 110k call) collects ~$380 net credit.

SPY May30: IV is subdued at 16%, but there were three large call prints in the 590 strike this morning. Someone is positioning for a breakout above the recent range.

ETH May31: front-month IV (72%) is unusually elevated vs Jun (68%). Short-dated IV is richly priced — good environment for theta collection.

What's your directional view and risk appetite?

3 — Monitor positions and alert on what matters

When a trading account is connected, the assistant watches your book around the clock. It checks position risk every 30 seconds against live market data and interrupts you only when something genuinely needs attention — not on every small move.

proactive alert — no question asked
Your short BTC 92,000 put (Jun27) is now delta −0.41 and BTC is down 4.2% in the last hour on above-average volume. The position has moved against you by ~$840. IV has risen 9 points — rolling the strike down to 85,000 would cost ~$210 and reduce your delta by half. Want me to navigate to the chain?

Conditions the assistant monitors:

4 — Navigate and control the interface

Say what you want to see, and the assistant takes you there. No need to click through exchange selectors — just describe it.

conversation example
show me the BTC December chain on Deribit and zoom to the ATM strikes
navigate_to(exchange="Deribit", underlying="BTC", expiry="Dec27")
set_zoom(strike_lo=85000, strike_hi=115000)
Done — showing Dec27 BTC, zoomed to ±15% around spot. The 100k strike has the highest open interest. ATM IV is sitting at 81%.

5 — Guide beginners step by step

For someone new to options, the assistant acts as a patient instructor that never judges the question. It explains what you are looking at in plain language, points to the relevant part of the screen, and builds intuition incrementally — like that trainer running alongside you.

guided learning example
I see the chain but I don't understand what I'm looking at
Let's start with the shape. Each row is a strike price — the price at which you could buy or sell the underlying. Calls (right column) let you profit if the price goes up. Puts (left column) let you profit if it goes down.

The thick line through the middle is the current price of the underlying. Options close to that line are "at the money" — they're the most actively traded. Options far from it are cheaper but less likely to pay off.

See the numbers in the bid and ask columns? The bid is what someone will pay you to sell the option. The ask is what it costs you to buy it. The gap between them is the cost of the round trip — the spread you give up every time you enter and exit.

Want to try understanding what the IV column means next?

Bridging prediction markets and options

Prediction markets let anyone bet on a plain-English question: "Will BTC hit $70k this week?" They are intuitive, popular, and growing — but they are binary, illiquid, and unregulated. You win or you lose. You cannot adjust the position, hedge it, or roll it forward.

Options on regulated exchanges are the professional equivalent of the same bet, with far more flexibility. The problem is translation: most people who understand the Polymarket question do not know which strike, which expiry, and which exchange to use — or how to compare the implied probability in the option price to the Polymarket odds.

option.watch already streams both Polymarket contracts and BTC/ETH option chains. The AI assistant closes the gap between them.

prediction market → options translation
Polymarket says 28% chance BTC hits 70k this week. Is that a good bet?
scan_market(exchange="Deribit", underlying="BTC", expiry="this_friday")
find_option(strike=70000, expiry="this_friday", type="call")
The BTC 70,000 call expiring Friday is bid 0.0008 / ask 0.0011 BTC on Deribit, which is roughly $105 per contract at spot. Black-Scholes implies about 19% probability of expiring in the money.

Polymarket is pricing the same event at 28%. That gap — 28% vs 19% — could mean the options market is cheap, or that Polymarket is over-pricing the move. Two ways to look at it:

If you trust Polymarket: buy the 70k call. You pay $105 for something the crowd thinks is worth $280 in expected value. Max loss is the premium.

If you trust the options market: sell the Polymarket YES contract. The options market, which prices this professionally with real hedging flow, says 19% — not 28%.

Either way the options route gives you something Polymarket doesn't: you can close early, roll to a different strike, or hedge delta if BTC moves toward 70k before Friday. Want me to show you the chain?

The AI can run this comparison for any prediction market question that maps to a price level or event — and it has live data from both sides to make the comparison real, not hypothetical.

Beyond arbitrage, this is a powerful onramp. Someone who only knows "I think BTC will be above 70k Friday" can be guided — step by step, in plain language — from that intuition to a specific, regulated, capital-efficient trade on a real exchange, with defined risk. That is the practical meaning of uniting prediction markets and options markets.

What it never does

The assistant never places, modifies, or cancels orders.

It has no write access to any trading account. It can read your positions (with explicit permission) and it can tell you what it thinks you should do — but you place every trade yourself. The boundary between suggestion and execution is absolute and by design. No autonomous trading, no "auto-pilot" modes, no API keys with trading permissions.
Capability Can do
Answer any options questionyes
Scan live markets across all exchangesyes
Read your open positions and P&Lyes, with permission
Proactively alert on risk eventsyes
Navigate the interface to a chain or strikeyes
Suggest specific trades with current pricesyes
Translate prediction market questions to options tradesyes
Compare Polymarket implied probability vs options marketyes
Place, modify, or cancel ordersnever
Access accounts on external exchangesnever
Trade autonomously on your behalfnever
Store conversation history on a third-party servernever

How it works

01
You type a question or the monitor fires an alert
Chat panel inside the option.watch workspace — same window, no new tab
02
The C++ server assembles live context
Current chain snapshot, IV curve, flow events, position data — injected directly into the prompt before the LLM sees your question
03
A local LLM answers — no cloud, no data leaving your network
Powered by Qwen3, running on-premises. Conversations are not sent to any external AI service
04
Tool calls execute server-side if needed
scan_market, find_option, navigate_to — the LLM can request actions; the server executes them and returns data for the next response turn
05
Response streams back to the chat panel
Token by token, same as any modern chat interface — no waiting for the full response

Privacy

The LLM runs locally. No conversation text, no market queries, and no position data are sent to any third-party AI provider. The assistant is an on-premises product: your questions, your positions, and your trading ideas stay on the option.watch server you connect to — and optionally self-hosted on your own infrastructure.

Availability

The AI assistant is in active development and will be available first as a beta to premium subscribers. If you want early access or have ideas for what the assistant should be able to do, send us a note.