stocks-llm
Server Details
Plain-English stock research on SEC filings, US indexes and UK FTSE 350. Informational, not advice.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-06-18
- URL
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 3 tools
The three tools target clearly distinct intents: thematic discovery (search_companies), single-entity deep lookup (get_company_profile), and a fixed two-ticker side-by-side (compare_companies). Overlap is minimal — search returns lightweight numerics but is explicitly a discovery/ranking operation, while profile and compare are entity-addressed. An agent can pick the right tool from the query shape alone.
All three follow a consistent snake_case verb_noun pattern: compare_companies, get_company_profile, search_companies. The verbs (compare/get/search) accurately signal the operation type, and no conventions are mixed.
Three tools is on the lean side, but each maps to a genuinely different access pattern and the single-profile tool absorbs a large amount of capability internally rather than fragmenting into a dozen sub-tools. Slightly thin for a data platform — there is no bulk or multi-ticker retrieval path — but well-scoped overall.
For a read-only informational research surface, the lifecycle is essentially covered: find companies by theme/filter/sort, inspect one company fully, and compare two. Minor gaps remain — comparison is hard-capped at two tickers with no way to batch a watchlist, no historical price series, and no universe/sector enumeration tool — but agents can work around these via repeated calls.
Available Tools
3 toolscompare_companiesAInspect
Head-to-head comparison of TWO companies (mirrors the /compare/{a}-vs-{b} page): the same real, dated side-by-side metrics — price, market cap, P/E, PEG, margins, ROE/ROA, debt/equity, revenue growth, dividend yield + streak, beta, 1-year return — plus each company's one-line pitch. A metric absent for either company is returned as null (never fabricated). Both tickers must be in the covered universe: US public companies: S&P 500 + NASDAQ-100 + Dow Jones + Russell 1000 + Russell 2000 (deduped — effectively the Russell 3000 plus the Dow), spanning large-, mid-, AND small-cap companies, plus the major US exchange-traded funds (ETFs). UK public companies: the FTSE 350 (FTSE 100 + FTSE 250), listed on the London Stock Exchange. Data: SEC EDGAR filings + market-data fundamentals, each figure dated. Informational only, NOT financial advice.
Inputs: ticker_a (required), ticker_b (required).
get_company_profileAInspect
Full enriched profile for ONE company (mirrors the /company/{ticker} page): name, ticker, sector, the LLM business summary + one-line pitch, key fundamentals (price + as-of date, market cap, margins, ROE/ROA, valuation), dividend facts, competitors, plus its latest SEC filings — the DATED evidence you cite. May also include insider_cluster_buying: a dated open-market cluster-buy DISCLOSURE fact (≥3 distinct insiders each buying, SEC code P, within a 30-day window; disclosed Form 4 activity, not a prediction or buy signal; only when one qualifies). May also include sector_percentiles: where the company's metrics (P/E, margins, ROE, growth, yield, absolute FCF) sit relative to covered peers in the same sector, each with a summary + peer count + as-of date (a relative position, not a valuation/quality judgment; only when rankable). May also include long_term_picture: business quality, price vs fundamentals, long-run trend and cash returned as x/10 from the published Top 10 pillars, with a summary sentence (a published formula, not advice). May also include valuation_context: today's P/E, P/S, P/FCF against the company's own fiscal-year range and a reverse DCF's implied free-cash-flow growth (a simple model, not a price target). May also include dividend_safety: a single-fiscal-year dividend-coverage assessment from SEC XBRL (earnings + free-cash-flow payout ratios; outcome supported/stretched/not_assessable/stale_dividend_data; past coverage, not a forecast; only when the company paid a dividend). stale_dividend_data means the latest fiscal year with a reported dividend payment is OLDER than the latest year of reported accounts (latest_financials_fiscal_year), so no ratio is given — do not infer from it that the dividend was cut, nor that the data is missing. May also include filing_changes: an AI-assisted, source-bounded comparison of the two most-recent same-form SEC filings (Risk Factors / MD&A additions, removals, wording changes) with each side's filed date + SEC url — a source comparison, not a materiality/legal/investment conclusion; only when a cached diff exists. May also include institutional_ownership: dated SEC Form 13F-HR holdings — the count of distinct investment managers reporting this security at the latest quarter-end plus a derived between-quarters position-change count (quarter-end snapshots filed within 45 days, not real-time; a reported holdings change, never a same-day trade; only when 13F data exists). May also include short_interest: reported FINRA equity short interest (shares) as of the latest settlement date, with the source-reported change + days-to-cover (delayed, bi-monthly; not real-time, not a percent of float, not a squeeze signal; only when short-interest data exists). May also include signal_agreement: how many standard chart readings are currently in a bullish vs bearish state for this company — bullish.count / bearish.count with the member signals as { key, note }, neutral, and computable, the number of signals we could actually compute for it. ALWAYS cite the count against computable ("6 of 9"): a company with a short price history has no 200-day average and is scored out of fewer signals, so a bare count is not comparable across companies. Describes past price behaviour only — never a buy/sell signal (only when computable signals exist). May also include next_earnings_date — which never travels alone: next_earnings_is_estimate says what KIND of date it is (1 = the data source's own ESTIMATE, 0 = a confirmed/announced date, null = unknown), and next_earnings_as_of says when it was last refreshed. State the kind whenever you cite the date — an estimated date is "expected around", never "reports on". Dates are refreshed in rotation, so a date at or before its next_earnings_as_of stamp may have moved since we read it; treat it as the date we hold rather than as settled. next_earnings_date is always today or later: a date that has already passed is withheld. next_earnings_status says which case you are in — "upcoming" means we hold a forward date; "passed" means our calendar's last expected date has gone by, and that date is carried in last_expected_earnings_date for context only. Never present last_expected_earnings_date as upcoming, and never read it as confirmation that the company reported — only last_earnings below is a filed fact. May also include last_earnings — the OPPOSITE KIND of date: the day the company filed its earnings announcement with the SEC (Form 8-K, Item 2.02). It is BACKWARD-looking and is a FILED FACT, never an estimate, so never present it as upcoming and never mix it with next_earnings_date without saying which is which. When its stale flag is true the announcement is over two quarters old — call it the most recent one IN OUR DATA rather than implying the company reported recently. When date is null, reason says why and the difference matters: foreign_private_issuer means the company reports on Form 6-K, which carries no earnings item code, so we hold no confirmed date — a coverage boundary, NOT a claim that it has not reported. Never substitute another filing or date for a missing one. Identify by ticker (preferred) or name. Returns an error object if the company is NOT in the covered universe. Universe: US public companies: S&P 500 + NASDAQ-100 + Dow Jones + Russell 1000 + Russell 2000 (deduped — effectively the Russell 3000 plus the Dow), spanning large-, mid-, AND small-cap companies, plus the major US exchange-traded funds (ETFs). UK public companies: the FTSE 350 (FTSE 100 + FTSE 250), listed on the London Stock Exchange. Data: SEC EDGAR filings + market-data fundamentals, each figure dated. Informational only, NOT financial advice.
Inputs: name, ticker.
search_companiesAInspect
Natural-language / thematic search over the stocks-llm knowledge base of US public companies: S&P 500 + NASDAQ-100 + Dow Jones + Russell 1000 + Russell 2000 (deduped — effectively the Russell 3000 plus the Dow), spanning large-, mid-, AND small-cap companies, plus the major US exchange-traded funds (ETFs). UK public companies: the FTSE 350 (FTSE 100 + FTSE 250), listed on the London Stock Exchange. Data: SEC EDGAR filings + market-data fundamentals, each figure dated. Informational only, NOT financial advice. Matches a plain-English THEME (e.g. "high-power semiconductor devices", "GLP-1 diabetes drugs") against each company's enriched business summary, products/technologies and theme tags — not just its sector. Returns real DB companies with dated performance numerics (price + price_as_of, market_cap, price_return_1y, revenue_growth_yoy, net_margin) and dividend facts. Use sort to RANK the matched set and the filters to narrow it.
Inputs: query (required), dir, sort, limit, +60 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Free US stock research from SEC filings: five plain answers on 3,000+ stocks. No API key.
Plain-English US stock market news and daily market analysis. Free, no login.
Primary-source equity research: scores, screens, filings, insider trades and 13F holdings.
Source-cited financial research on US and Nordic company filings.
Related MCP Servers
- FlicenseNot gradedqualityCmaintenanceEnables risk analysis of US public companies by analyzing 8-K filings and insider activity using live SEC EDGAR data.-
- FlicenseNot gradedqualityDmaintenanceExplains why US stocks moved. AI analysis for 'why did Tesla drop?' queries. Covers S\&P 500, NASDAQ 100, Dow 30 (~550 stocks).1-
- AlicenseNot gradedqualityCmaintenanceEnables researching one US or KRX stock at a time from an AI chat client, returning SEC EDGAR financials and insider filings alongside unofficial daily price history and roughly 40 computed technical, risk, entry-scenario and EPS-scenario factors. It is read-only, holds no brokerage connection, and labels every value with its source and as-of date.MIT
- AlicenseAqualityBmaintenancePre-computed financial market intelligence for AI agents. Stocks, crypto, and ETFs.978 npm5MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.