LITMUS
Technical signals ยท AI agent debates

HOW IT WORKS

THE SIGNALS

Type any US stock symbol above and Litmus pulls its last year of daily bars, testing it against its 20-day, 50-day and 200-day moving averages โ€” pure price-vs-average rules, no black box. STRONG BUY = the stock reclaimed its key averages. BUY = it's retesting a still-rising average and holding. Every signal card shows entry reference, hard stop-loss, invalidation level, and position size for a $10k account risking 1%.

THE LITMUS SCORE (0โ€“100)

Every ticker gets a Litmus Score built on @thesetupfactory's setup framework โ€” six measurable components, highest score at the top of every list:

70+ is green, 50โ€“69 amber, under 50 red. Hover any score for its full breakdown. Missing inputs score neutral โ€” never faked.

Alongside it, MOMO is the classic momentum score (v2): 3/6/12-month momentum, 52-week-high proximity, RS vs S&P, trend stack, and volume. LITMUS times the entry; MOMO finds the leaders. The best setups score well on both.

AGENT DEBATES

Tap โš”๏ธ Debate the agents on any card and a team of AI trading agents goes to work: four analysts research fundamentals, technicals, sentiment and news in parallel; then a bull agent and a bear agent debate the ticker round by round; a trader agent makes the call; and a risk team (aggressive, neutral, conservative voices) stress-tests the trade before the final verdict. You get an instant health check right away; the full debate report lands in about two minutes and is cached for the day.

CREDIT

The debate engine is based on TradingAgents, an open-source multi-agent LLM trading framework by researchers at UC Berkeley (TauricResearch/TradingAgents). This site runs a hosted build of the framework (abinashsa/TradingAgents).

DATA

Daily bars: Yahoo Finance's public chart API, fetched live at scan time through this site's own Cloudflare Worker proxy. Not financial advice โ€” for education and entertainment only.