Drillr — The financial MCP for AI agents
Financial data and research MCP for US/CN/JP equities: filings, statements, ownership, signals.
Links
README
From the repo.
🌐 English · 中文
drillr · The financial data and research MCP for AI agents
Filings, statements, earnings, ownership, events, executives, analyst data, company discovery and research signals for US, China and Japan equities — every figure traced to the document it came from.
Browser sign-in, no API key to copy. One hosted Streamable HTTP endpoint, ten tools: turn a name or a description into tickers, list and search a company's filings (structured as-reported facts and the filed text), run read-only SQL over the financial tables, and pull research signals from earnings calls and news.
⭐ If drillr helps your agent, star us — that's how we know to keep building this in the open.
Quick Start
- Sign up at drillr.ai (80 free credits, no card)
- Add
https://gateway.drillr.ai/mcp/datato an OAuth-capable MCP client asdrillr-data - Approve drillr Data in the browser. Restart the client — MCP tools load at startup.
Per-client instructions, kept current on the docs site: https://drillr.ai/developer/mcp-install.md. Or paste this into any coding agent with a shell:
Please install the drillr MCP server for me and walk me through the sign-in: server drillr-data, url https://gateway.drillr.ai/mcp/data, auth browser OAuth. Use my client's own MCP command to add it. Per-client details: https://drillr.ai/developer/mcp-install.md
Claude Code
claude mcp add --scope user --transport http drillr-data https://gateway.drillr.ai/mcp/data
Restart, run /mcp, pick drillr-data → Authenticate.
Codex CLI
codex mcp add drillr-data --url https://gateway.drillr.ai/mcp/data
Codex starts the browser sign-in from this command. codex mcp login drillr-data re-runs it later.
Cursor / VS Code / Claude Desktop / other OAuth hosts
{ "mcpServers": { "drillr-data": { "type": "http", "url": "https://gateway.drillr.ai/mcp/data" } } }
Cursor reads .cursor/mcp.json; VS Code reads .vscode/mcp.json with a top-level servers key; Claude Desktop takes the URL under Settings → Connectors → Add custom connector.
API-key fallback
For hosts without browser OAuth (Hermes, OpenClaw, self-built clients) or for REST: create an external key at drillr.ai/account/api-keys, keep it in a secret store, and add it as a bearer header on the same URL. Never configure OAuth and a static header on the same server entry.
"headers": { "Authorization": "Bearer <YOUR_DRILLR_API_KEY>" }
# Hermes Agent
mcp_servers:
drillr-data:
url: 'https://gateway.drillr.ai/mcp/data'
headers: { Authorization: 'Bearer <YOUR_DRILLR_API_KEY>' }
Agent skill
The companion skill teaches an agent how to onboard a user and use drillr well over REST or MCP, with a generated endpoint reference:
npx skills add Little-Grebe-Inc/drillr-skill
https://github.com/Little-Grebe-Inc/drillr-skill
Smithery / Claude Code plugin fallbacks
npx -y @smithery/cli install drillr/drillr --client claude # Smithery: leave the key empty on OAuth clients
/plugin marketplace add Little-Grebe-Inc/drillr-mcp-server # Claude Code plugin marketplace
/plugin install drillr
Hello World
"What does NVDA's latest 10-K say about supply commitments, and how did data-center revenue move over the last four quarters?"
- The agent calls
ticker_lookupif it only has a name,filing_listto see what is indexed, thenfiling_search— which returns as-reported facts (value, period, XBRL concept, accession number) and the passages they sit in, in one call. - For the quarterly series it calls
run_sqlonfinancial_statements. - You get an answer whose every number cites its filing. Typical round-trip 8–15 s.
The ten tools
Each tool has a page at https://drillr.ai/docs/mcp/<tool> with parameters, limits and error shapes; the server's tools/list carries the same descriptions. The set stays small on purpose: the agent composes them.
| Group | Tool | What it does | Cost |
|---|---|---|---|
| Company | ticker_lookup | Name, brand or ticker substring → canonical symbols, historical names included | free |
company_search | Natural-language description → companies with a match reason; US / CN / JP / HK / KR; all matches returned | 3–5 cr | |
| Filings | filing_list | Which filings are indexed for a ticker: fiscal period, type, filing date | 0.1 cr |
filing_search | Search one company's filings; returns structured as-reported facts and the filing text together | 0.1 cr | |
| Datasets | list_tables | Alt-data category index, or the tables under up to five categories | free |
get_table_schema | One table's columns, types and usage note — required filters, coverage, gotchas | free | |
run_sql | One read-only PostgreSQL SELECT over the financial, market and alt-data tables | 0.1 cr | |
| Signal | industry_inflections | Industry changes synthesized from many US earnings calls: mechanism, scope, degree, per-company impact | 1 cr |
ai_adoption | Concrete enterprise AI applications disclosed on US earnings calls: workflow, stage, value, evidence | 1 cr | |
news_search | Semantic search over company and market news: storylines, events, attributed claims | 0.2 cr |
What the data is
Built by drillr from primary sources — filings taken from each market's official channel (SEC EDGAR, cninfo, EDINET) and parsed in house, what companies publish on the web (IR pages, earnings calls, news), and drillr's own estimates. No data vendor in between. Every reported figure links back to the passage it was filed in. Provenance per dataset: https://drillr.ai/docs/provenance.
| Dataset | What is in it | How to reach it | Coverage |
|---|---|---|---|
| Company | Ticker resolution from any name, code, ISIN, CIK or CUSIP; discovery by description; profile (exchange, industry, listing date, website, headcount) | ticker_lookup, company_search; company_snapshot via SQL | US CN JP (discovery + HK KR) |
| Filings | Filing index with form type, date and official link; full-text search in EN / ZH / JA; as-reported fact records with value, unit, period, filing and position, linked in a knowledge graph so the right version of a figure wins | filing_list, filing_search | US CN JP |
| Financials | Income statement, balance sheet, cash flow as reported under each market's standard, annual and quarterly; precomputed valuation, margin, return, growth, leverage and liquidity metrics | run_sql on financial_statements, company_snapshot | US JP HK CN KR |
| Prices | Daily / weekly / monthly OHLCV, indices, pre- and after-market quotes | run_sql on price_volume_history, index_price, equity_extended_rt | equities US JP HK CN KR; indices, FX, crypto, commodities |
| Earnings | Calendar with estimates and actuals; structured call summaries (highlights, guidance, risks, segments, Q&A) | run_sql on earning_call_calendar, earning_call_summary | US JP |
| Ownership | Insider holdings and transactions (Forms 3 / 4 / 5); institutional quarterly positions (13F-HR) | run_sql on insider_and_institution_activities | US |
| Events | 8-K corporate events: executive changes, deals, debt issuance, securities offerings; 13D / G stake changes | run_sql on executive_change, company_deal_events, debt_issuance, securities_offering | US |
| Executives | Roster with status; annual compensation from DEF 14A | run_sql on executive_profile, executive_compensation | US |
| Analyst | Individual rating and target changes; consensus distribution and targets | run_sql on analyst_ratings, analyst_ratings_consensus | US |
| Signal | Conclusions drawn by drillr with the evidence quoted: industry inflections, enterprise AI adoption, cross-source news storylines and attributed claims | industry_inflections, ai_adoption, news_search | US (news: US CN JP + macro) |
| Alt-data (secondary) | 9 categories, 65 tables around the AI supply chain and macro: energy & power, data centers, semiconductors, compute pricing, model development, inference economics, macro & trade, prediction markets, critical minerals | list_tables → get_table_schema → run_sql | global |
Ticker forms: US bare (AAPL), A-shares .SH / .SZ (600519.SH), Japan .T (6758.T), Hong Kong .HK (00700.HK), Korea .KS / .KQ (005930.KS); indices ^GSPC; quote symbols with . or ^ in SQL. Freshness: filings, ownership and events within minutes of the filing; calls and news within minutes to hours; statements and analyst data daily.
Benchmark: the drillr MCP-powered model scores 94.06 % on Vals Finance Agent, and drillr publishes its own restatement-aware benchmark — https://drillr.ai/drillr-benchmark.
REST API
The same data as 29 typed REST endpoints for scripts and pipelines — one X-API-KEY header, JSON out, public OpenAPI 3.1 contract:
curl -H "X-API-KEY: $DRILLR_API_KEY" \
"https://gateway.drillr.ai/api/v2/income-statements?ticker=AAPL&period=FY&limit=4"
Reference: https://drillr.ai/docs/api · Contract: https://gateway.drillr.ai/api/v2/openapi.json. The older /api/v1/data/* tool mirrors are documented in docs/rest-api.md.
Pricing and limits
One credit wallet across MCP and REST. Free: 80 credits on sign-up. Plus $29 / 300 cr per month, Ultra $99 / 1,500 cr, Enterprise for redistribution and SLA. run_sql returns 100 rows (500 on Ultra), 10 s per statement, 5 in flight; company_search is capped per day (20 / 100 / 500). Failed calls are not billed. https://drillr.ai/pricing
Out of scope
Private or unlisted companies · on-chain crypto metrics (CEX prices only) · options chains, order book, tick data · placing orders · drillr does not produce its own price forecasts.
Community
| Discord | |
| Devs building agentic research products — office hours, debugging help, early access. | Chinese-speaking dev community — fastest product feedback. |
License
MIT — see LICENSE.
Config for your environment
Use the endpoint URL below in your config. No API key — you connect directly.
Tool
OS
Config file: ~/.cursor/mcp.json
{
"mcpServers": {
"mcp-server": {
"url": "https://gateway.drillr.ai/mcp/data"
}
}
}Paste into mcpServers in the config file. Restart Cursor after saving.
If this MCP is also published on mcpchannel.ai, you can subscribe from Browse and use the gateway config there instead.