Run a fundamentals strategy in one minute.

Take this one: insiders buying, institutions adding, fundamentals holding up. Three conditions across Form 4, 13F and 10-Q — one agent run over the whole library, and only one company passed.

Three filters,
run over everything.

No hard thresholds. Who is buying, how serious the money is, who is adding, whether the filings hold up — the agent judges all of it.

Find US stocks where insiders are buying, institutions are adding, and fundamentals look good

▾Agent: querying AkkruDatavia MCP / API✓

Candidates (Form 4 screen)	Insiders buying	Institutions adding	Fundamentals
NCLH	✓ $28.1M	✓ 4 straight quarters	✓ agent pass
FISV	✓ $1.7M	✗ holders −32% in 4 qtrs	—
ADSK	✓ $1.8M	✗ flat	—
RSG · NDAQ · others	10% owners / small buys	—	—

✓ Agent: 7 names from the first screen, ruled out one by one — some were routine top-holder buying, some institutional adds were just broker pass-through, and FISV had executives buying while institutions sold. One left — NCLH: the CEO and directors paying out of pocket, with Elliott and Baupost both opening new positions.

An agent analyst
reads the filing.

This is where judgment comes in — the agent reads the Q1 10-Q (filed May 4) and gives an analyst-style assessment. Click any number for the formula and source.

▾Agent: querying AkkruDatavia MCP / API✓

NCLH · Q1 10-Q (filed 05-04)	Value	Agent's read
Revenue growth YoY	+9.6% Growth = (Revenue − Revenue_prior) / |Revenue_prior| Revenue (Q1 2026)2,331.2M Revenue (Q1 2025)2,127.6M	Accelerating (prior qtr +2.9%)
D/E (period end)	5.75 D/E = (LT debt + ST debt + commercial paper) / Equity Period-end valueas reported 10-QFiled 2026-05-04	Deleveraging (6.22 → 5.75)
P/E @ filing-date price $17.20	14.2× P/E = Price / EPS Price$17.20 · 2026-05-04 EPS10-Q source	Not pricing in the improvement
Free cash flow (TTM)	−$0.95B FCF = CFO − Capex TTM rollforwardQuarter-by-quarter 10-QFiled 2026-05-04	Shipbuilding capex; still negative

✓ Agent's assessment: a classic improving, heavily levered balance sheet — growth accelerating, leverage falling, and a 14× multiple that isn't pricing the improvement in. The negative FCF comes from a shipbuilding capex cycle with a defined end, not a deteriorating business. The filing supports the bet — pass. Not one hard threshold in the whole strategy; all three steps are judgment. That's the difference between Agent × API and a pure quant screen.

How the signal
played out.

Replayed by disclosure date — each row is the day the information became public. Hold until the next filing.

▾Agent: querying AkkruDatavia MCP / API✓

Disclosed	What became public	Price	Held to Aug 3
05-15	Q1 13F: Elliott opens a 13.19M-share position, Baupost opens 3.63M	~$15	≈ +32%
05-19	Form 4: director Cil buys $0.15M	$14.91	+32.5%
05-26	Form 4: CEO Chidsey buys $2.5M	$16.37	+20.7%
06-03	Form 4: director Pagliuca buys $25M — signal complete	~$18.1	+9.2%
08-03	Q2 10-Q filed — information reset, window ends	$19.76	—

✓ Agent: buy on any disclosure date, hold to the next filing — every entry positive (+9% to +33%). After the Q2 filing cut guidance, the stock gave back 22% in a month; that's a new decision under new information. The discipline: enter on disclosure, reassess at the next filing.

A technical demonstration of the AkkruData API. All data comes from public SEC filings; all dates are disclosure dates. A single period and a single name are not statistical evidence, and "hold to the next filing" is a hypothesis to be tested. Not investment advice, and not an offer or recommendation to buy or sell any security. 13F disclosures lag by up to 45 days; insider and institutional activity does not predict future performance. akkrudata.ai