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

1. Sign up at drillr.ai (80 free credits, no card)
2. Add https://gateway.drillr.ai/mcp/data to an OAuth-capable MCP client as drillr-data
3. 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?"

1. The agent calls ticker_lookup if it only has a name, filing_list to see what is indexed, then filing_search — which returns as-reported facts (value, period, XBRL concept, accession number) and the passages they sit in, in one call.
2. For the quarterly series it calls run_sql on financial_statements.
3. 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.

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.