10 Free AI Prompts That Turn Any LLM Into a Crypto Trading Analyst

Why most crypto AI prompts fail

Search “AI prompts for crypto trading” and you get a hundred listicles of clever-sounding instructions. Paste one into ChatGPT and you quickly hit the wall: the model confidently analyzes a market it can’t see. It doesn’t know the current funding rate, today’s open interest, which coins whales are positioned in, or what regime the market is in right now. It’s pattern-matching on training data that’s months stale.

A prompt is only as good as the data behind it. So we built the missing half: a free, open AI Prompt Library where every prompt is wired to a live CryptoDataAPI endpoint. You fetch real numbers, paste them into the prompt’s {data} slot, and the model reasons over what the market is actually doing. Below are 10 to start with — copy-paste, works with Claude, GPT-4o and Gemini.

The 10 prompts

Each prompt names the exact endpoint and fields it consumes and returns a structured, decision-ready output. Free-tier prompts work with a free API key; a few pull Pro feeds. Open any of them in the prompt library to copy the full text.

1. Funding Rate Extremes Scanner

Surfaces coins where perpetual funding has gone extreme — crowded, over-leveraged positioning that often precedes a squeeze. This is the prompt to monitor Hyperliquid funding rates people keep asking for. Data: /api/v1/derivatives/funding-rates.

2. Market Regime Detection

Reads our quant HMM engine’s current regime and probability distribution and translates it into a plain-English, risk-appropriate playbook — a real prompt to detect a crypto market regime with an LLM. Data: /api/v1/quant/market.

3. Open Interest Divergence Scanner

Finds coins where price and open interest disagree — new-money trends vs. hollow short-covering bounces. The prompt to scan crypto open interest divergence. Data: /api/v1/liquidity/oi-divergence.

4. Whale Positioning Monitor

Reads aggregate Hyperliquid whale (≥$100k accounts) long/short bias so you can see what large, informed perp traders are doing. The Claude prompt for crypto whale activity. Data: /api/v1/quant/whales.

5. Autonomous Portfolio Risk Monitor

An agent prompt that emits a single GREEN / AMBER / RED risk level from the confluence of regime, liquidity fragility and liquidations — the backbone of crypto trading automation. Data: /api/v1/quant/market + /liquidity/regime/score + /market-intelligence/liquidations.

6. Multi-Factor Signal Generator

Scans the whole coin universe and emits a ranked, evidence-backed watch shortlist by combining each coin’s regime with funding — one of the best prompts for building a crypto trading bot. Data: /api/v1/quant/coins + /derivatives/funding-rates.

7. MCP Market Analyst

Wire CryptoDataAPI into Claude with one command and this prompt makes the model pull a full one-call market snapshot itself — no copy-paste. Data: /api/v1/daily via the MCP server.

8. Telegram Alert Agent

Turns on-chain exchange-flow spikes and sentiment extremes into concise, push-ready Telegram alerts. Data: /api/v1/on-chain/exchange-flows/spike-alerts + /sentiment/fear-greed.

9. Volatility-Aware Position Sizer

Sizes a position inversely to expected volatility so risk-per-trade stays constant across coins — sizing math, not advice. Data: /api/v1/quant/coins/risk.

10. Regime-Aware Execution Controller

Adapts a bot’s execution style — order type, aggression, slice size — to the current regime and live order-book depth. Data: /api/v1/quant/market + /liquidity/depth.

How to prompt an LLM with live crypto data

The pattern is the same for every prompt: fetch the endpoint, drop the JSON into the {data} placeholder, send it to your model. Here’s the funding-rate example end to end — this is how to prompt an LLM to analyze crypto derivatives data:

# 1. Pull live data (free API key)
curl -H "X-API-Key: cdk_live_yourkey" \
  https://cryptodataapi.com/api/v1/derivatives/funding-rates

# 2. Paste the JSON into the prompt's {data} slot, then send to your LLM.

Prefer a native tool? Skip the copy-paste entirely and connect your model to live data with the MCP server — one command, then the model fetches for itself:

claude mcp add cryptodataapi -- npx -y cryptodataapi-mcp

Works in Claude Desktop, Cursor, Continue, or any MCP client. Most feeds run on a free API key; the quant regime, per-coin and whale feeds are on Pro.

Get the full library free

All 14 prompts (10 above plus backtesting and execution prompts) live in the free, open library — browse and copy them at cryptodataapi.com/prompts, or grab the markdown in the GitHub repo (PRs welcome). New prompts drop every week.