Prerequisites
Cryptohopper MCP configured in an MCP client — see the setup overview.
Explorer tier or higher. Technical analysis requires historical candle data, which is not available on Pioneer — see subscription tiers.
Familiarity with basic TA concepts (trend, RSI, MACD, moving averages) — see a practical guide to candles.
Setup steps
Open your MCP client
Issue the core TA prompt
A good response gives you a two- or three-paragraph summary covering trend, momentum, key levels, and any red flags.
Using the Cryptohopper MCP, run a technical analysis on ETH/USDT on Binance.
Data to pull
- 1h candles, last 150 bars
- 4h candles, last 150 bars
For each timeframe, compute
- RSI(14)
- MACD(12, 26, 9)
- The 20 / 50 / 200 EMAs
Analysis
- Identify the main trend on each timeframe (up, down, ranging).
- Identify recent support and resistance levels from candle highs and lows.
- Flag any divergence between RSI and price in the last 20 bars.
Note whether the timeframes agree on trend direction.
Output format: A brief written analysis — no tables.Save the prompt as a reusable template
Parameterize the pair and exchange: "Run a TA on {pair} on {exchange} using the pattern in my saved TA prompt." Most MCP clients support saved prompts or snippet libraries.
Expand the analysis for specific workflows
Add one of the extensions below depending on your use case.
Optional extensions
Use case | Add to the prompt |
Grid-bot prep | "Suggest a grid range based on the 4h high/low over the last 7 days and ATR(14). Output the range, step count, and per-grid size for a $5,000 position." See how to build a grid-bot parameter generator. |
Entry timing | "Based on the RSI and MACD readings, do you see a clear entry signal, or would you wait for a clearer setup?" |
Position sizing | "Compute realised volatility over the last 30 1h bars. Suggest a stop-loss distance based on 1.5× ATR(14)." |
Cost profile
Action | Explorer / Adventurer | Hero |
1h candles (150 bars) | 5 units | 1 unit |
4h candles (150 bars) | 5 units | 1 unit |
Total per TA run | ~10 units | ~2 units |
Explorer's 30,000 calls/week quota allows up to 3,000 TA runs per week. See rate limits explained.
Troubleshooting
The analysis feels generic or repetitive
The model is not using the candle data meaningfully. Be more specific — demand exact values: "report the RSI(14) reading on the last closed bar and the previous 5 bars" and "identify the highest high and lowest low in the last 50 bars." Vague prompts produce vague output; numeric prompts produce numeric output.
HISTORY_LIMIT_EXCEEDED
Your candle lookback is too deep for your tier. Reduce to 100–150 bars for most indicators. Pioneer does not support historical candles at all. See subscription tiers and candle tool reference.
Indicator values look wrong
The model computes indicators in its reasoning and can make arithmetic errors on long series. Ask it to show raw last-20-candle closes alongside the RSI value — if the closes look correct but RSI does not, add "show your RSI calculation step by step" to the prompt. Accuracy is generally good on lookbacks of ≤150 bars.
The analysis is based on the currently-open bar
Indicators on an open bar are noisy. Add to the prompt: "base all indicators on closed bars only; ignore the last (current) bar." See the open-vs-closed distinction in the candle tool reference.
Multi-timeframe disagreement confuses the output
A 1h showing bullish and a 4h showing bearish is a real and useful signal — but only if the prompt asks for it explicitly. Add: "explicitly state whether timeframes agree or disagree, and what the disagreement implies."
You want TA delivered on a schedule
See how to schedule Cryptohopper MCP workflows for automation and how to send MCP reports to Telegram, Discord, or email for delivery.
You want TA on many tokens at once
Cost scales linearly — a 20-token run is 20× a single-token run (~200 call units on Explorer, easily affordable). Watch your candle lookback: pulling 1,000 bars per token is the fastest way to exhaust quota.