Compare used-good prices across 9 Japanese shops + Rakuten Ichiba, MAFF data, prize giveaways.
Runs on your machine from its package; the steps are below.
At most: only reads or looks up information.
Odel charges nothing per call. The publisher may charge for its own service.
Runs locally over stdio
This server isn't hosted — your MCP client launches it from a package registry. Use one of the commands below, or drop the config into your client (e.g. Claude Desktop).
Distributed as an MCP bundle (`https://github.com/atushi1841/japan-market-mcp/releases/download/v0.1.0/japan-market-mcp.mcpb`) — download it and import the bundle in your MCP client.
An MCP server that lets AI agents (Claude, Cursor, VS Code Copilot, ChatGPT) compare used-good prices across leading Japanese shops in a single tool call. Instead of scraping site by site, an agent queries one server and gets inter-store price gaps for the same product — perfect for reseller arbitrage, price monitoring, and market research on the Japanese second-hand market.
The server wraps market-scraping and official-API actors (Rakuten Ichiba included) behind one MCP endpoint. Each tool returns price, brand, shop, condition and a link per item, with the shop-pair crossed so you see the price spread at a glance.
| Tool | Market | Shop pair |
|---|---|---|
search_camera_market | Used cameras & lenses | Kitamura + Fujiya Camera |
search_watch_market | Luxury watches | Jackroad + Kitamura used watches |
search_luxury_market | Used luxury brands | Komehyo + Jackroad |
search_instrument_market | Used instruments | Digimart + Ishibashi U-BOX |
search_offmall_market | General used goods | OffMall (Hard Off official, 800+ stores) |
search_kakaku_prices | New price comparison | Kakaku.com (aggregated, thousands of shops) |
search_car_market | Used cars | goo-net (nationwide, by body type) |
search_car_price_stats | Used car price stats (min/max/avg/median JPY) | goo-net (by model keyword, e.g. N-BOX, Alphard) |
search_japan_prize_giveaways | Current JP prize giveaways with deadlines (懸賞/プレゼント) | kenshou.club + cp.meikan.org |
get_japan_prize_giveaway_stats | Aggregate giveaway stats (counts, active, winners) | kenshou.club + cp.meikan.org |
search_rakuten_items | New retail products (official API) | Rakuten Ichiba 楽天市場 (keyword, price range, sort) |
get_rakuten_ranking | Best-seller ranking (official API) | Rakuten Ichiba 楽天市場 (overall or by genre) |
get_maff_market_report | Fresh produce wholesale market report (official MAFF open data) | Japan central markets (旬別, JPY/100kg, YoY%) |
search_maff_price_trend | Fresh produce wholesale price trend (official MAFF open data) | Multi-market, multi-period comparison |
get_maff_top_movers | Fresh produce YoY price movers (official MAFF open data) | Japan central markets (supply-shock signal) |
get_maff_markets | List available MAFF wholesale markets | no charge |
Each tool takes a keyword (e.g. SONY α7, ROLEX, Hermes, Fender) and returns results from the crossed shops. Leaving the keyword empty scans the full category.
Pay per event — $0.001 per search + actor start fee. Runs finish in seconds, so a full keyword lookup typically costs well under $0.01.
Run it from the Apify Store (Run button), then point your MCP client at the endpoint. It stays in standby mode.
MCP client config (Claude Desktop / Cursor / VS Code):
{
"mcpServers": {
"japan-market-mcp": {
"url": "https://fruitful-quintessence--japan-market-mcp.apify.actor/mcp",
"headers": {
"Authorization": "Bearer YOUR_APIFY_TOKEN"
}
}
}
}
<your-username>--japan-market-mcp.apify.actor/mcpAuthorization: Bearer header (Apify Console → Settings → Integrations)URL note: use a hyphen in the username (
fruitful-quintessence), not an underscore — an underscore hostname will not resolve.
llms-install.md — grounded guidance for AI agents (Cline, Claude Code, Cursor) that set up this server for a user..well-known/mcp.json advertises the endpoint, auth, tool list, and pricing in the standard MCP Directory discovery format.cline/mcp-marketplace — once approved it becomes one-click installable from Cline's marketplace.fruitful_quintessence/japan-market-mcp (pay-per-event).This MCP server connects to 86+ Apify Actors in the Kensho ecosystem for real-time data. Key related actors include:
# If GitHub API returns 403, use curl + git clone
curl -s https://raw.githubusercontent.com/atushi1841/japan-market-mcp/main/README.md > README.patch
# edit README.patch as needed
git clone https://github.com/atushi1841/japan-market-mcp.git
cd japan-market-mcp
cp ../README.patch README.md
git add README.md
git commit -m "update README via external run"
git push
In your claude_desktop_config.json:
{
"mcpServers": {
"japan-market-mcp": {
"command": "python",
"args": ["-m", "src.main"]
}
}
}
# Install dependencies
pip install -r requirements.txt
# Run as HTTP server on port 3000
APIFY_TOKEN=your_token python -m src.main --http --port 3000
# Test the endpoint
curl http://localhost:3000/mcp/tools | jq .
from src.main import mcp
# FastMCP instance can be embedded directly
Use Apify's Webhook node to POST JSON-RPC to POST /mcp endpoint.
docker build -t japan-market-mcp .
docker run -p 3000:3000 -e APIFY_TOKEN=your_token japan-market-mcp
"Compare used prices for Sony a7 III between Kitamura and Fujiya"
→ search_camera_market(keyword="SONY α7 III")
"What's the inter-store price gap for a Rolex Submariner?"
→ search_watch_market(keyword="ROLEX Submariner")
"What's the market price for a Hermes Birkin 25?"
→ search_luxury_market(keyword="Hermes Birkin 25")
"What's a used Fender Stratocaster going for?"
→ search_instrument_market(keyword="Fender Stratocaster")
"Show this 旬's wholesale market report for Tokyo (Toyosu) fresh vegetables"
→ get_maff_market_report(market="toyosu")
"Which fresh produce items have spiked in price vs last year at Osaka?"
→ get_maff_top_movers(market="oosakaho")
"Track the price of daikon (だいこん) across Tokyo and Osaka this month"
→ search_maff_price_trend(item="だいこん", markets=["toyosu","oosakaho"])
robots.txt.pip install -r requirements.txt
APIFY_TOKEN=apify_api_xxx python -m src.main
# Streamable HTTP endpoint at http://localhost:3000/mcp
MIT
Connect this MCP server to your AI client (Claude Desktop, Cursor, VS Code) in one command:
npx @smithery/cli install atushi1841/japan-market-mcp --client claude
Replace claude with cursor, vscode, or cline for other clients.
Alternatively, install directly from the Smithery registry.
Note: Smithery server listing is pending verification. Once verified, this server will appear in search results with useCount tracking.