Skip to content

jquants-mcp

Talk through Japanese stocks with Claude — backed by the J-Quants API v2.

jquants-mcp is a Japanese-equities-focused MCP (Model Context Protocol) server. It gives Claude — Desktop, CLI, or mobile — 47 specialist tools and a local SQLite cache, turning it into a hands-on companion for your stock research rather than a one-off query tool.

What you can ask Claude

Once jquants-mcp is connected, conversational queries like these just work:

  • "Sector performance ranking today" — sector-by-sector advance/decline
  • "Kioxia (285A) — 3-month candlestick chart" — split-adjusted with SMA overlays
  • "Q4 earnings digest for the five major trading houses" — pulls the latest fins_summary rows for each
  • "Stocks hitting new year-to-date highs" — detect_ytd_high_low screener
  • "What's the code for SoftBank?" — reverse-lookup via search_equities
  • "Compare TOPIX vs Nikkei 225 over 1 year" — multi-stock comparison chart
  • "Show me the sector briefing" — advance/decline count, AD ratio, top/bottom 5 sectors by return, and cheapest-PER sector ranking, all in one call

Features

  • Ask in plain English or Japanese — jquants-mcp picks the right tool and returns a clean answer. No tool names to memorise.
  • Instant responses — market data is cached locally, so most queries never hit the network at all.
  • Works on any J-Quants plan — Free through Premium, auto-detected.
  • Chains naturally — "screen for top movers, then chart the leader" works as a single request.

Get started

  • Quickstart → — install, register your API key, and have Claude answer your first stock question in 5 minutes.
  • Tools → — the user-facing tour of what jquants-mcp can do.
  • FAQ → — plan recommendations, common errors, and tips.

For the full technical reference (config schema, deployment shapes, multi-user mode, OAuth setup, every tool with parameter tables), see the README on GitHub.

Not investment advice

This software is a data-access tool, not a financial advisory service. Investment decisions are made at your own risk. See the full disclaimer for details.