MCP + REST API

Put the whole US-market database
into your AI.

Ask in plain language and the agent queries first-hand data with 49 tools, then answers with sources. From raw SEC text to options Greeks, in one place.

You

Deep dive on NVDA: latest-quarter statement highlights, whether the valuation looks stretched, whether institutions and insiders are adding or trimming, news this past week, then a verdict.

The agent calls
get_data_coverage()
get_objective_report(ticker="NVDA")
get_company_news(ticker="NVDA", days=7)
get_analysis(ticker="NVDA")
Claude

A dated NVDA brief that separates first-party SEC facts from vendor news.

Connect in three steps

Works with Claude Desktop, Claude Code, Cursor, or any MCP-speaking agent; or build on the REST API.

1

Connect the MCP server

One command adds Stockfacts to your agent.

claude mcp add stockfacts
2

Get an API key

Sign up free, paste the key into your agent config to authorize.

Get a free API key
3

Start asking

Ask in natural language; the agent handles tool calls and queries.

You can ask these

Seventeen data domains and 49 tools in one query layer. Every domain answers plain-language questions.

SEC filings, full text

"What are the risk factors in AAPL's latest 10-K?"

Three financial statements

"NVDA's gross margin trend over the last four quarters?"

Insider Form 4

"Which CEOs bought their own stock this month?"

13F institutional holdings

"Who did Berkshire add last quarter?"

Options chain + Greeks

"Any IV skew on TSLA calls expiring next week?"

Earnings-call transcripts

"What did MSFT say about AI capex on the call?"

News sentiment

"Is AMD's news sentiment positive or negative this week?"

Daily / hourly prices

"META's price action over the past year?"

Earnings calendar

"Which large caps report next week?"

Dividends / splits

"AAPL's dividend and split history?"

Macro / commodities / indices

"Latest CPI, 10-year Treasury yield and VIX?"

ETF holdings / reverse lookup

"Which ETFs hold NVDA?"

Six common use cases

Each is a ready-to-copy prompt, plus the real tools the agent calls and what comes back.

One-shot deep-dive on a stock

Let the agent build an evidence-backed write-up from first-party SEC financials, ownership and news, instead of reciting stale training data.

You ask

Deep dive on NVDA: latest-quarter statement highlights, whether the valuation looks stretched, whether institutions and insiders are adding or trimming, news this past week, then a verdict.

The agent calls
get_data_coverageget_objective_reportget_company_newsget_analysis
You get

A dated NVDA brief that separates first-party SEC facts from vendor news.

Market-wide insider-buying scan

Instead of one ticker, sweep the whole market for executives buying their own stock on the open market.

You ask

Which companies had a CEO or CFO buy shares on the open market in the past 30 days? Show a few with dollar amounts.

The agent calls
screen_insider_buysget_companylist_insider_trades
You get

A dollar-ranked insider-buying shortlist that drills into any name's Form 4 detail.

Track this quarter's institutional flows

Read first-party 13F to see who entered or exited a name, or flip it to inspect a whale's whole book.

You ask

Which institutions sharply added AAPL last quarter and which trimmed? Also list Berkshire's top ten holdings for the quarter.

The agent calls
list_13f_top_buyerslist_13f_top_sellerssearch_institutionslist_13f_portfolio
You get

Add and trim leaderboards for AAPL plus Berkshire's holdings by market value, all flagged with the 45-day filing lag.

Earnings-week and premarket scan

Have the agent map who reports next week and who is moving hardest today, so you prep ahead.

You ask

Which large companies report between July 13 and 17 next week? Then show today's biggest gainers, losers and most-active names, and whether any tie to earnings or news.

The agent calls
get_earnings_calendarget_market_moversget_market_news
You get

A next-week earnings calendar joined to today's movers and the news behind them.

Ask the earnings-call transcript

Ask what management said about demand and guidance on the latest call, quoted verbatim, then cross-check against the quarter's actual numbers.

You ask

On MSFT's latest earnings call, what did the CEO say about AI and cloud capex? Summarize, quote the exact words, and cross-check against the quarter's actual figures.

The agent calls
get_earnings_transcriptget_latest_period
You get

A management-tone summary with verbatim quotes, tagged by quarter and speaker, checked against the quarter's financials.

Free-form SQL across the universe

Pro

Pro runs read-only SELECT for cross-company stats and custom screens no training data can answer; or wire the data into your own system via the REST API.

You ask

Use SQL to find the 15 semiconductor names with the highest latest-quarter gross margin, listing ticker, gross profit and revenue.

The agent calls
describe_tableexecute_readonly_sql
Example SQL
SELECT c.ticker, i.gross_profit, i.revenue
FROM income_statements i
JOIN companies c ON c.id = i.company_id
WHERE c.industry ILIKE '%semiconductor%'
ORDER BY i.gross_profit::numeric / NULLIF(i.revenue, 0) DESC
LIMIT 15;
You get

A cross-company ranked table, run through a read-only role with a 5-second statement timeout, safe and auditable.

Or use the REST API
curl -H "Authorization: Bearer $STOCKFACTS_KEY" \
  "https://trading-agent-api.zeabur.app/api/screener/insider-buys?since_days=90&limit=20"

Grab a key and let your agent start querying US stocks.

The data is free to read. You sign up only to get your own API key, connect MCP, or raise your limits.