A practical, copy-paste workflow for using altFINS preset filters through the altFINS MCP server and Claude Desktop — with three live trade setups pulled from today’s market.

Most crypto traders waste hours clicking through screener menus, memorizing preset codes, and copy-pasting tickers between tabs. With the altFINS Model Context Protocol (MCP) server connected to Claude AI, that workflow collapses to a single conversation. You describe the trade setup you want; Claude picks the right preset filter from altFINS’s 150+ curated strategies, runs it against live market data, and hands back a ranked, decision-ready shortlist.

This guide shows the exact workflow, the preset categories that produce the highest signal-to-noise, and three real examples generated while writing this article.

Why Preset Filters Beat Manual Screening

A preset filter is a pre-built combination of indicators that defines a specific technical setup. Instead of manually configuring RSI thresholds, trend scores, MACD crossovers, and volume conditions, you call a single preset that encodes the entire recipe.

— We have explained preset filters in this article.

You can use any of these directly in the altFINS Crypto Screener. The breakthrough with MCP is that you no longer need to remember the preset codes. You describe the setup, and Claude maps it to the right filter.

The Three-Step Workflow

This pattern works for any preset on the platform. It keeps you focused on trade ideas instead of filter mechanics.

Step 1 — Describe the Setup, Not the Filter

Tell Claude what kind of trade you’re hunting for in trader language. Examples: “oversold coins that are still in a confirmed uptrend”, “resistance breakouts with above-average volume”, “DeFi tokens with growing TVL”. Skip the preset codes. Claude handles the translation.

Step 2 — Specify the Columns That Drive the Decision

A ticker list alone is useless. Ask for the metrics you actually use to evaluate a trade: market cap (to filter illiquid names), short and medium-term trend scores, RSI14, recent price change, and relative volume. These five columns alone separate clean setups from noise.

Step 3 — Pick the Winners

Ask Claude to highlight the cleanest setups and flag the obvious passes. The model is good at spotting when a “breakout” is actually still in a downtrend, or when an “oversold” coin is dying rather than dip-buyable.

Three Live Examples From Today’s Market

All data below was pulled live from the altFINS MCP server. Use it to see how the workflow plays out across three completely different setup categories.

Example 1: Oversold Coins in a Confirmed Uptrend

This is the textbook “buy the pullback” setup: a coin in a strong medium-term uptrend that has corrected hard enough to flash oversold on short-term RSI. The preset returned only three names — and that scarcity is itself a useful signal about current market conditions.

Symbol	Market Cap	RSI14	1W	1M	Short Trend	Medium Trend
BLUAI	$11.9M	39.5	-28%	-12%	Down (4/10)	Strong Up (9/10)
FAR	$11.3M	35.5	-20%	-9%	Down (2/10)	Up (7/10)
AFC	$7.9M	32.6	-31%	-8%	Strong Down (1/10)	Strong Up (10/10)

Every row shares the same structure: short-term trend has rolled over, medium-term trend is still intact. That is the pullback-buy setup in its purest form. Whether you’d actually trade these is a separate decision — all three are sub-$15M caps with real liquidity risk — but the preset surfaced exactly the structure it promised.

Example 2: Resistance Breakouts With Confirming Volume

Different idea: coins that just punched through a resistance level, ideally with volume backing the move. Adding Relative Volume to the requested columns separates conviction breakouts from suspicious ones.

Symbol	Market Cap	RSI14	1W	Rel. Volume	Trend
NEAR	$2.17B	71.4	+11%	1.27x	Strong Up (10/10)
HTX	$1.81B	79.5	+3%	1.10x	Strong Up (10/10)
2Z	$338M	67.9	+10%	2.61x	Up (8/10)
LIT	$310M	74.0	+54%	1.77x	Strong Up (10/10)
TRAC	$212M	69.4	+37%	2.75x	Strong Up (10/10)

Triage in seconds: NEAR and HTX broke out on modest volume (1.1x to 1.3x average) — the move is real but unenthusiastic. 2Z, LIT, and TRAC ran 1.8x to 2.8x normal volume, the kind of conviction that tends to extend rather than fade. LIT and TRAC are already RSI-extended, so a pullback entry beats a chase entry.

That’s the entire screening process: one preset, one column choice, 90 seconds of decision-making.

Example 3: Pure Oversold (And Why It’s a Trap Without Context)

For contrast, here’s what happens when you screen on RSI alone with no trend filter. The preset returns 100+ names — and almost every one is in a Strong Down (0/10) trend on both timeframes.

Symbol	Market Cap	RSI14	1W	1M	Trends
BCH	$7.4B	27.6	-13%	-15%	Strong Down / Strong Down
TRUMP	$485M	27.9	-13%	-28%	Strong Down / Strong Down
CHIP	$105M	3.2	-93%	-99.96%	Strong Down / Strong Down
TORN	$30M	28.3	-11%	-30%	Strong Down / Strong Down

The takeaway: a low RSI is not a buy signal on its own. Most oversold coins are oversold because they’re trending down — and trending instruments can stay oversold for weeks.

Ready-to-Use Prompts

Copy any of these directly into Claude with the altFINS MCP connected. Each one targets a different setup type.

For more advanced multi-filter prompts, see the altFINS guide on 8 AI-Powered Trade Strategy Prompts Using altFINS MCP.

How to Set It Up

Three steps to get the preset workflow running:

1. Create an altFINS account and generate an API key from your dashboard.
2. Add the altFINS MCP server to Claude Desktop using the configuration in the official connection guide.
3. Quit Claude Desktop, open it again, and start prompting in plain English.

   

See MCP documentation
See Pre-set filter documentation

If you prefer staying inside the altFINS platform, the altFINS AI Copilot offers the same natural-language screener experience natively, without any external setup.

Three Honest Limitations

A preset is a starting filter, not a thesis. Every screen surfaces tokens that look great on paper and are untradeable in practice — illiquid, manipulated, or fundamentally weak. Treat output as a watchlist, never as a buy list.

Timeframe matters and is often invisible in the preset name. Most presets default to the daily candle. Scalpers on 15-minute charts need different presets and different display types. The MCP integration accepts a timeInterval parameter — use it.

Verify before retrying. If a prompt returns nothing, ask Claude to confirm the preset name exists in altFINS before retrying with variations. Hallucinated preset codes are the most common failure mode.

The Bigger Picture

The interesting shift here isn’t speed. Speed is the surface benefit. The real change is in what kind of question you can practically ask a screener.

Before MCP: “let me click through ten presets and see what catches my eye.” After MCP: “give me oversold names in confirmed uptrends, ranked by liquidity, with trend scores and RSI displayed, and flag anything under $20M market cap.”

That’s a different relationship with the tool. You stay in idea-space. The screener handles translation. The trade ideas surface faster, cleaner, and with the context you need to act on them.

Disclaimer: Examples in this article illustrate the workflow only — they are not trade recommendations. Crypto markets are volatile and small-cap tokens carry substantial liquidity and counterparty risk. Always conduct your own research before trading.