AI Prompts for Crypto Trading

You are a sceptical crypto research analyst. Using web search, write a one-page brief on [PROJECT]: what it does, who uses it, how it earns revenue, the role of its token, main competitors and the three biggest risks. Cite a source and a date for every number.

Verify: Open the cited sources for revenue and user figures. Drop any number that has no date.

Give me the strongest bull case and the strongest bear case for [TOKEN] over the next [TIMEFRAME]. For each, list the three facts it depends on and the data I could check to confirm or reject each fact.

Verify: Both cases are hypotheses. Check the listed facts yourself before acting on either.

Which narratives are driving the [SECTOR] sector right now? For each, name two or three representative tokens, the catalyst behind it, and the signs that would show the narrative is fading. Use sources from the last 30 days only.

Verify: Check the dates on every source. Search results can be weeks old while looking current.

List scheduled events for [TOKEN] in the next 60 days: token unlocks, upgrades or mainnet dates, governance votes, exchange listings, and macro events that usually move crypto. Give the date, a source link and expected impact as low, medium or high.

Verify: Confirm each date on the project's official channels or an unlock tracker. Dates shift.

From the attached whitepaper and docs, build a table of total supply, circulating supply, allocation by group (team, investors, treasury, community), cliff dates and vesting schedule. Add the page number for every figure.

Verify: Compare circulating supply with a block explorer or a data aggregator. Docs go out of date.

Using this vesting schedule: [PASTE], estimate how many tokens unlock each month for the next 12 months, as a percentage of current circulating supply. Flag any month above 5% and show your working.

Verify: Recalculate one or two months by hand. Models slip on multi-step arithmetic.

Read this whitepaper as a sceptical auditor. List vague claims, missing technical detail, unrealistic yields, centralised control points such as admin keys or upgradeable contracts, and anything that contradicts itself. Quote the passage for each point.

Verify: Check admin and upgrade rights in the audit report or the verified contract, not the whitepaper.

Explain in plain English how value reaches holders of [TOKEN]: fees, burns, staking rewards, governance rights. Then say whether the protocol would work just as well without the token, and why.

Verify: Confirm fee and burn mechanics in official docs or on-chain, not in blog posts.

Using [DATA SOURCE], show net flows into [TOKEN] from wallets labelled smart money or funds over the last 7 and 30 days. List the five largest wallets involved and whether each is accumulating or distributing.

Verify: Open two or three of the wallets on a block explorer and check the labels make sense.

Using [DATA SOURCE], compare [COIN] exchange inflows and outflows over the last 30 days with the 90-day average. Explain what the change usually signals and the common exceptions to that reading.

Verify: Exchange wallet labels differ by provider. Compare with a second source before trusting a trend.

Analyse the top 100 holders of [TOKEN] on [CHAIN]. Separate exchanges, contracts, team wallets and unknown wallets. What share of supply do the top 10 non-exchange wallets hold?

Verify: Use a holder-cluster map such as Bubblemaps to see whether "different" wallets are linked.

Write a Dune SQL query that returns daily active addresses and daily transfer volume for [TOKEN CONTRACT] on [CHAIN] for the last 90 days. Explain each table and filter you use.

Verify: Run it and spot-check one day against a block explorer. AI-written SQL often filters the wrong table.

Search X for posts about [TOKEN] in the last 24 hours. Show the ten most-engaged posts with account names, follower counts and account age, then summarise the tone. Do not summarise without showing the posts first.

Verify: Open the posts. Check whether the engagement comes from many accounts or from a handful.

Of the accounts posting about [TOKEN] today, roughly what share were created in the last 90 days or post almost only about this token? Does the pattern look organic or coordinated? Explain the evidence.

Verify: A model can only estimate this. Treat a "coordinated" verdict as a reason to dig, not proof.

Using [SENTIMENT SOURCE], compare social volume and sentiment for [TOKEN] with its price over the last 90 days. Did sentiment spikes come before, during or after the big price moves?

Verify: Look at the chart yourself. Ninety days is a small sample for any correlation.

Sentiment on [TOKEN] looks extremely [POSITIVE OR NEGATIVE]. Give historical crypto examples where similar extremes marked a top or bottom, and examples where they did not. What would make this time different?

Verify: Check the historical examples exist. Models sometimes invent convenient precedents.

Here is a [TIMEFRAME] chart of [PAIR]. Describe the trend, market structure, key support and resistance zones and volume behaviour. Say clearly which values you cannot read from the image.

Verify: Confirm levels on the live chart. Image reading is approximate and axes are often misread.

My readings for [PAIR]. Weekly: [NOTES]. Daily: [NOTES]. 4-hour: [NOTES]. Tell me where the timeframes agree and where they conflict, and what a patient trader would wait for before entering.

Verify: The answer is only as good as your notes. Recheck each reading before relying on it.

Explain what [INDICATOR] measures, how it is calculated, common settings used on crypto charts, and the market conditions in which it gives false signals. Keep it under 250 words.

Verify: Check the formula against your charting platform's documentation.

My trade idea: [LONG OR SHORT] [PAIR] because [REASON]. Before discussing entries, describe the price action that would prove me wrong and where a logical stop would sit based on market structure.

Verify: Place the stop where the chart says, not where the model estimated it.

Turn this idea into exact rules a computer can follow: [IDEA]. Include entry, exit, stop-loss, take-profit, position size and timeframe. Then list every assumption you had to make.

Verify: Each assumption is a decision. Change the ones you disagree with before testing.

Write a backtest specification for these rules: [RULES]. Include data source, a date range covering a bull, a bear and a sideways phase, fees of [FEE]% per side, slippage, and the metrics to report: return, max drawdown, win rate, profit factor and number of trades.

Verify: Make sure the test period includes at least one crash.

Write a [PINE SCRIPT OR FREQTRADE PYTHON] strategy for these rules: [RULES]. No look-ahead bias, fees included, every line commented. Then list three ways the code could still be wrong.

Verify: Read the code and test it on a short period where you can check each trade by hand.

Here are my backtest results and the parameters I optimised: [PASTE]. How likely is this to be overfit? Suggest an out-of-sample and walk-forward test plan with dates.

Verify: Tell the model how many variants you tried. It cannot judge overfitting without that number.

Account size [SIZE]. Risk per trade 1%. Entry [ENTRY], stop [STOP], fees [FEE]% per side. Calculate the position size in coins and dollars, show the working, and the loss if the stop slips by a further 1%.

Verify: Recheck the arithmetic with a calculator before placing the order.

I want to open a [LEVERAGE]x long on [PAIR] at [ENTRY] with isolated margin. Estimate the liquidation price, explain how funding payments affect the position over a week, and what price move wipes out the margin.

Verify: Use the exchange's own liquidation calculator. Formulas and maintenance margins differ by venue.

My portfolio: [HOLDINGS WITH PERCENTAGES]. Estimate the drawdown if BTC fell 30% and altcoins moved with their typical sensitivity to BTC. Which positions dominate the risk?

Verify: This is a rough scenario, not a forecast. Altcoins can fall far more than history suggests.

Write a one-page set of risk rules for a part-time crypto trader with [SIZE] capital: maximum risk per trade, maximum daily and weekly loss, maximum open positions, leverage limits, and what to do after three losses in a row.

Verify: Adjust to your own tolerance, then set the limits in your bot or exchange account.

Review these grid bot settings: pair [PAIR], range [LOW] to [HIGH], [N] grids, investment [SIZE], fee [FEE]% per side. Calculate profit per grid after fees, and explain what happens if price leaves the range on either side.

Verify: Cross-check profit per grid in a grid bot calculator. Thin grids often lose to fees.

Is [PAIR] suitable for a grid bot right now? Assess the last 30 to 90 days of range behaviour, volatility and trend strength, suggest a range with reasons, and say what data you used.

Verify: Confirm the range on the chart. Past ranges break, and trends are what hurt grid bots.

Review this DCA bot: base order [X], [N] safety orders at [STEP]% steps with a [SCALE]x volume scale, take profit [TP]%. How much capital is needed if every safety order fills, and how far must price fall to fill them all?

Verify: Recalculate total capital yourself. Safety-order maths compounds faster than it looks.

I am connecting a trading bot to [EXCHANGE]. List the safest API key settings: permissions, IP whitelisting, sub-account setup, key rotation, and what to do if the key leaks.

Verify: Follow the exchange's own API docs, since menu names vary. Withdrawals stay disabled, always.

Here are my trades this week as CSV, with my notes: [PASTE]. Find my best and worst setups, every rule I broke, and the single change most likely to improve results. Be blunt.

Verify: Spot-check the statistics on a few rows. Models miscount in long tables.

This trade lost [AMOUNT]: [DETAILS]. Separate what was bad luck from what was a bad decision. Which rule, if I had followed it, would have changed the outcome?

Verify: The model only sees your version of events. Be honest in the details.

Compare my bot's live results: [PASTE] with its paper-trading results: [PASTE]. Explain the gap: fees, slippage, timing, market regime or a difference in logic.

Verify: Check actual fills in the exchange history before blaming the market.

Create a trade journal template with fields for setup, thesis, entry, stop, target, size, emotions, outcome and lesson, plus a weekly summary section with spreadsheet formulas for win rate and average R multiple.

Verify: Test the formulas on a few sample rows before relying on them.

Check [PLATFORM NAME] for signs of an investment scam: regulator warnings from the FCA, CFTC, SEC or state regulators, company registration, who runs it, how deposits and withdrawals work, and complaints about withdrawals. Cite sources.

Verify: Search the FCA Warning List and your national regulator yourself. Absence of a warning is not approval.

This platform claims [CLAIM]. Show what that return compounds to over one year, whether any legitimate strategy has delivered it consistently, and what regulators say about such claims.

Verify: Guaranteed or fixed returns are a red flag regulators name explicitly. No further checks needed.

For a token at contract [ADDRESS] on [CHAIN], what should I check before buying: verified source code, mint and blacklist functions, liquidity lock, holder concentration and transfer taxes? Name a tool for each check.

Verify: Get contract addresses only from the official project site, never from a chat answer.

I received this message offering crypto trading help or an airdrop: [PASTE]. List every red flag, what the sender probably wants next, and how I can verify whether it is genuine.

Verify: Never share seed phrases, keys or remote access, whatever any answer says.