Drillr — The financial MCP for AI agents
@Little-Grebe-Inc
Drillr — The financial MCP for AI agents について
Drillr — the financial MCP for AI agents. Standardized financial data, SEC filings semantic search, live signals, and 24 alt-data categories.
基本情報
設定
以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。
{
"mcpServers": {
"drillr": {
"type": "http",
"url": "https://gateway.drillr.ai/mcp/data",
"headers": {
"Authorization": "Bearer ${DRILLR_API_KEY}"
}
}
}
}ツール
8Use to discover which SEC filings exist for a ticker before searching content. For the actual content use sec_report_search instead. List indexed SEC filings for a given ticker with a summary header. Returns: summary (period coverage, per-type counts) + table of up to 50 filings (fiscal_year, fiscal_quarter, filing_type, filing_date, period_start, period_end). filing_types filter: omit for main reports only (10-K, 10-Q, 20-F, S-1, DEF 14A and /A amendments; excludes 8-K/6-K); pass [] for all indexed types; pass explicit allowlist to override.
Use when you need narrative content (risk factors, MD&A, guidance language). For consolidated financial numbers use run_sql on financial_statements instead. SEC report semantic search. Search paragraph content in 10-K/10-Q/20-F/6-K filings. Search Strategy: - (1) Time-specific queries ("recent", "last year", "Q4 2025", etc.): Apply a time range filter upfront to narrow results to the relevant period. - (2) Broad queries with no time requirement: Broad first, narrow later. First 1-2 searches: use MINIMAL filters. Do NOT set filing_types or period unless explicitly needed. This discovers what information exists in what filings. Later searches: add period/type filters to narrow down. Filing types: 10-K (US annual), 10-Q (US quarterly), 8-K (US current/material events), 20-F (foreign annual), 6-K (foreign current), S-1 (US IPO registration), F-1 (foreign IPO registration). SCOPE: - (1) Indexes ONLY company-filed SEC reports. Does NOT index institutional filings (13F-HR, 13D, 13G). For institutional holdings, use insider_and_institution_activities with source='institution'. - (2) For standardized consolidated financial numbers (revenue, margins, EPS, etc.), use financial_statements first — SEC filings contain narrative context, not structured data. High-Value Search Scenarios (use proactively when relevant): - Share dilution / SBC / buyback → search "Shareholders' Equity" or "Capital Stock" for share-level (not just dollar-level) breakdowns - Facilities / properties → search "Properties" for owned vs leased status, lease terms, land use rights - Interest rate sensitivity → search "Market Risk" or "Interest Rate Risk" for floating-rate debt exposure, swap arrangements - Segment revenue breakdown → search "Segment Information" or "Revenue Disaggregation" - Risk factors → search "Risk Factors" for specific risks - Management guidance → search "Outlook" or "Guidance" in MD&A - Accounting policies → search "Critical Accounting Policies" - Stock split history → search "Stock Split" for split ratios and dates Use for: segment revenue, risk factors, management discussion, guidance, accounting policies, operational metrics, related party transactions, property details, share structure.
Use for qualitative company discovery (industry, business model, supply chain, competitors, management background). For numerical screening (revenue, margins, ratios, growth rates) use run_sql on company_snapshot instead. Drillr's company knowledge base — searchable across industry classification, product offerings, business model, segment structure, competitive landscape, supply chain, management background, and customer profile. Pass a natural language description (e.g. "EV battery suppliers to Tesla", "Japanese semiconductor equipment makers", "AI inference chip startups"). Returns a structured list of matching companies with context snippets. ONLY for finding a LIST of companies by description.
Use when you need structured rows from financial/market/altdata tables. For SEC filing narrative use sec_report_search instead. For company qualitative discovery use company_search instead. For recent news + market events use signal_list instead. Execute a read-only PostgreSQL SELECT query on financial and alternative-data tables. Call `get_table_schema` first to look up columns for a specific table. For alternative-data categories, call `list_tables(categories=[...])` to see tables + columns. SQL constraints: - No CTE (`WITH ... AS`) — use subqueries instead. - Date columns are TEXT — use plain string comparison (`period_end >= '2024-01'`). Never use `::date` cast or `INTERVAL` arithmetic. - No `ROUND(float8, int)` — use `CAST(value AS DECIMAL(10,2))` if rounding is needed. - Structured-data queries: always filter by ticker (`WHERE ticker IN ('A','B','C')`; screening: `ticker NOT LIKE '%-%'` to exclude preferred stock). Alternative data is macro/industry/patent — no ticker filter required. Structured data (tables grouped by domain): ## Market Data - price_volume_history: Historical OHLCV price data, ~32M rows. MUST filter by ticker AND time_frame ('daily'|'weekly'|'monthly') to avoid timeout. Columns: ticker, period_end, time_frame, open, high, low, close, adj_close, volume, turnover, vwap, change, change_percent. Ticker conventions: * STOCK / ETF — bare 1–5 letters (AAPL, MSFT, SPY, QQQ, VOO, TLT). Foreign listings use exchange suffix: 1557.T (JP), 310960.KS (KR). * INDEX — leading "^" (^GSPC=S&P 500, ^DJI, ^IXIC=NASDAQ Composite, ^NDX=NASDAQ 100, ^RUT=Russell 2000, ^VIX). Foreign indices: ^FTSE, ^GDAXI, ^N225, 000001.SS (Shanghai Composite), FTSEMIB.MI. * COMMODITY — [code]+USD or +USX (CLUSD=WTI futures, BZUSD=Brent, NGUSD=natgas, GCUSD=gold, SIUSD=silver, HGUSD=copper, ZCUSX=corn, CTUSX=cotton). Suffix USX = priced in cents. * FOREX — 6 letters base+quote, no separator (EURUSD, USDJPY, GBPUSD, USDCNY). * CRYPTO — [token]+USD (BTCUSD, ETHUSD, SOLUSD, DOGEUSD, USDTUSD, USDCUSD). Pitfalls: 1. Same asset, different tickers: NASDAQ 100 → index ^NDX (~26000) vs ETF QQQ (~640). Pick the one matching user intent. 2. WTI spot ≠ futures. CLUSD here is NYMEX futures, NOT spot (FRED has spot via DCOILWTICO). 3. Tickers with "." or "^" MUST be quoted in SQL: `WHERE ticker = '^NDX'`. - index_price: Real-time price snapshot for 3 major US indices: S&P 500 (^GSPC), Dow Jones (^DJI), NASDAQ 100 (^NDX). - index_composition: Index constituent membership for S&P 500 / Dow Jones / NASDAQ 100 / Nikkei 225 (^N225) / TOPIX (^TPX). Includes current and historical members (active rows have exit_date IS NULL). - equity_extended_rt: Extended-hours snapshot, one row per ticker (~6.1K US equities; PK=ticker; refreshed every few minutes — see extended_updated_at). Use ONLY for pre-market / after-hours / overnight quotes. Columns: ticker, company_name, market, price_current (last regular-session close), then three parallel blocks pre_* / after_* / overnight_* (price, change_val, change_rate, high, low, volume, turnover), plus extended_updated_at. For regular-session OHLCV history use price_volume_history; for valuation / fundamentals use company_snapshot. Japan tickers are NOT in this view yet — extended-hours data is US-only. ## Fundamentals - financial_statements: Quarterly and annual financial data covering income statement, balance sheet, and cash flow statement. Filter by ticker + fiscal_period ('FY' annual, 'Q1'..'Q4' quarterly). - company_snapshot: Real-time company snapshot, one row per company. Pre-computed metrics (ratios, percentages, per-share figures, growth rates) only — NOT raw statement line items. For raw statements use financial_statements; for qualitative discovery use company_search. ## Earnings - earning_call_summary: Earnings call data with AI-structured summaries (management_highlights, guidance, risks, segment_performance, q_and_a). Filter by ticker + period_end (yyyy-mm). NOT for structured financial numbers. - earning_call_calendar: Earnings conference call schedule with EPS and revenue estimates vs actuals. `date` column is the scheduled earnings call datetime (UTC timestamptz) — NOT the press-release / 10-Q filing date (the filing typically lands the same day or shortly after). Use eps_actual IS NULL for upcoming calls, IS NOT NULL for reported. ## Analyst Coverage - analyst_ratings: Individual analyst rating events — every re-rating is a separate row. ~565K rows, 519 firms. Filter by ticker; `date` is TEXT (string comparison, no ::date cast). importance >= 4 for high-impact calls. - analyst_ratings_consensus: Per-ticker analyst consensus rollup — one row per ticker, refreshed daily. Columns: strong_buy/buy/hold/sell/strong_sell counts, total_analysts, consensus, pt_consensus/high/low. ## Ownership & Insider Activity - insider_and_institution_activities: Insider trades (Form 3/4/5) and institutional holding changes (13D/G/F, 13F-HR) from SEC EDGAR. Filter by ticker + source ('insider' vs 'institution'). For institutions, filer_name is the holder; shares/market_value give position size. ## Corporate Events (8-K) - executive_change: 8-K executive change events — appointments, departures, reasons, successors, effective dates. Filter by ticker + event_type. - company_deal_events: Single source of truth for corporate deal activity (8-K 1.01/1.02/2.01, S-4, press releases). M&A, financing, material agreements as event-based rows (announced/signed/amended/closed/terminated/waived). Filter by ticker only — do NOT filter by deal_type. One deal may span multiple rows linked via deal_fingerprint. - debt_issuance: 8-K debt issuance events — principal amount, interest rate, maturity, lender, use of proceeds. Use is_current = true for latest version when amendments exist. - securities_offering: 8-K securities offering events — shares, price per share, total proceeds, investors, lock-up. Use is_current = true for latest version when amendments exist. ## Executive Profiles & Compensation - executive_profile: Key executive profiles — name, title, compensation, gender, birth year. One row per executive per company. - executive_compensation: Annual executive compensation breakdown — salary, bonus, stock awards, option awards, total. Filter by ticker + year. ## Alternative Data For alternative-data tables, call list_tables(categories=[...]) first.
Use BEFORE run_sql when you're unsure which columns a table has. Look up column definitions (name, type, description) for a data table.
Use to convert between fiscal year/quarter and calendar months for a given ticker. Companies have different fiscal year starts (Apple Sep, Nvidia Jan) — call this before filtering on period_end columns. Fiscal year ↔ calendar month bidirectional conversion. Forward: ticker + fiscal_year + fiscal_quarter → period_start/period_end. Reverse: ticker + yyyy_mm → fiscal_year/fiscal_quarter.
List alternative-data tables under the given categories. Returns each table's name, one-line purpose, and column names (call get_table_schema if you need column types/comments). Batch up to 5 categories in one call. Use this BEFORE run_sql when you want to explore alt-data — run_sql alone won't tell you which tables exist. Available categories: - Energy & Power — US power plants, electricity prices, regional hourly generation/demand - Data Centers — facilities, GPU clusters, cooling - Semiconductors — AI chip specs, sales, ownership, foundry revenue, customs trade - Compute Pricing — GPU rental, cloud VM spot/on-demand, instance specs - Model Development — model specs, benchmarks, AI companies, AI polling, LLM arena - Inference Economics — LLM API pricing across providers - Macro & Trade — UN Comtrade, US Census trade flows, FRED macro series - Prediction Markets — Polymarket and Kalshi events, markets, trades, daily aggregates - Critical Minerals — USGS mineral deposits, country supply, critical materials
Use when you need recent news, events, or market-moving signals for specific tickers or sectors. For SEC filing narrative use sec_report_search instead. Recent news + market events filtered by ticker / sector / time range. Each row is one signal: id, headline, summary, suggested_tickers, sector, score, trigger_sources, earliest_trigger_event_time, created_at, tags. Continuously updating feed. Coverage: - ~6,900 tickers across US + ADRs of global companies - Cross-asset: equities, macro, geopolitics, commodities, crypto - Default sort by earliest_trigger_event_time DESC Parameters: - tickers (optional): array of tickers — returns signals whose suggested_tickers overlaps any of these - sector (optional): array of sector strings — returns signals whose sector overlaps any of these - from_date (optional): ISO 8601 timestamp; filter earliest_trigger_event_time >= from_date - to_date (optional): ISO 8601 timestamp; filter earliest_trigger_event_time <= to_date - order_by (optional, default earliest_trigger_event_time): 'created_at' | 'earliest_trigger_event_time' - limit (optional, default 20, max 100): max results - offset (optional, default 0): pagination offset
概要
What is Drillr?
Drillr is a financial MCP server for AI agents, providing a research data backend that covers global equities, fundamentals, SEC filings, earnings, markets, analyst coverage, news, signals, and alt-data. It is designed for developers building agentic financial research workflows.
How to use Drillr?
Sign up at drillr.ai, obtain an external-scope API key (format drl_xxxxxxxx_xxx..., 45 characters), then configure the server in your MCP host’s mcp.json using Streamable HTTP transport with a Bearer token header. Options include manual config, Smithery one-line install, or the Claude Code plugin.
Key features of Drillr
- 9 tools covering financial data, SEC filings, company search, and signals
- Standardized financial data over 90+ tables (statements, ratios, earnings, insider, ownership)
- Paragraph-level semantic search over SEC filings (10-K, 10-Q, 20-F, etc.)
- Live cross-asset signal feed across ~6,900 tickers
- Ontology-based company discovery by business model, supply chain, or theme
- Global equities coverage: US and Japan (Hong Kong, A-shares, Korea coming soon)
- Alt-data for the AI value chain (energy, data centers, semiconductors, LLM pricing, etc.)
Use cases of Drillr
- Pull and compare financial statements (e.g., gross margin) for multiple companies
- Search SEC filings for specific disclosures with cited paragraphs
- Discover companies by qualitative description (e.g., “EV battery suppliers to Tesla”)
- Monitor live signals across equities, macro, geopolitics, commodities, and crypto
- Resolve company names or ticker substrings to canonical tickers
FAQ from Drillr
What data does Drillr cover?
Global equities (US and Japan), fundamentals back to the 1980s, SEC filings with paragraph-level search, earnings transcripts with AI summaries, markets (equities, ETFs, indices, forex, crypto, commodities), analyst coverage, news and signals, and alt-data for the AI value chain.
What is out of scope for Drillr?
Private/unlisted companies, on-chain crypto metrics (only CEX prices), options chains, real-time order book, intraday tick data, retail brokerage actions, and price forecasts (Drillr surfaces analyst consensus but does not produce its own forecasts).
How do I authenticate with Drillr?
Use a Bearer token in the HTTP header. The API key starts with drl_ and is 45 characters long. Set it in the Authorization header as Bearer <YOUR_DRILLR_API_KEY>.
What transport does Drillr use?
Drillr uses Streamable HTTP transport. It is compatible with any MCP host that supports this transport (Claude Code, Cursor, VS Code, Hermes Agent, OpenClaw, ChatGPT MCP, etc.).
Is there a REST API for Drillr?
Yes, every MCP tool has a 1:1 REST endpoint. The same API key and billing apply. REST responses include a _credits envelope on every 2xx response.
「金融とコマース」の他のコンテンツ
Crypto Fear & Greed Index MCP Server
kukapayProviding real-time and historical Crypto Fear & Greed Index data
Financial Modeling Prep MCP (Model Context Protocol)
imbenrabiA Model Context Protocol (MCP) implementation for Financial Modeling Prep, enabling AI assistants to access and analyze financial data, stock information, company fundamentals, and market insights.
Market Fiyatı MCP Server
EnesCinr
Crypto Price & Market Analysis MCP Server
truss44A Model Context Protocol (MCP) server that provides real-time cryptocurrency analysis via CoinCap's API. Enables Claude and other MCP clients to fetch crypto prices, analyze market trends, and track historical data.
MCP Yahoo Finance
maxscheijenA Model Context Protocol (MCP) server for Yahoo Finance.
コメント