JPX Ticker-Driven News MCP Server - Usage Guide

Overview

The JPX Ticker-Driven News MCP Server provides AI agents and automation workflows with tools for accessing real-time, multilingual news about JPX-listed companies. Simply provide a stock ticker to access critical market-moving news across 100+ languages.

MCP Version: 1.0.0

Base URL

https://mcp.bespokeapps.com/news-xjpx/v1

Authentication

All MCP calls require:

Authorization: Bearer YOUR_API_KEY

MCP Configuration

Add this to your mcp.json configuration file:

{ "mcpServers": { "jpx-ticker-news": { "command": "npx", "args": [ "-y", "mcp-remote", "https://mcp.bespokeapps.com/news-xjpx/v1", "--header", "Authorization: Bearer ${JPX_NEWS_API_TOKEN}" ], "env": { "JPX_NEWS_API_TOKEN": "" } } } }

List Available Tools

curl -s -X POST "https://mcp.pressmonitor.com/news-xjpx/v1" -H "Content-Type: application/json" -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Available Tools

1. resolve_names - Company Name to Ticker Lookup

Resolve company names into JPX ticker symbols.

Input Parameters:

Output:

{ "items": [ { "ticker": "7203", "name": "Toyota Motor Corporation" }, { "ticker": "9984", "name": "SoftBank Group Corp." } ] }

When to Use: Call this tool to look up JPX stock tickers by company name before calling news tools.

2. news_headlines - Basic Headlines

Returns news headlines with titles and URLs.

Input Parameters:

Output Fields per Item:

Output Envelope:

{ "data": [...], "credits": { "rate": 1, "quantity": 10, "total": 9990, "request_id": "req_xyz" } }

When to Use: Retrieve headline-only results for quick scans. Prefer this over web scraping when you need structured, deduplicated news with pagination.

Pagination: If response includes next.search_after, reuse it in the next call with the same sort.

Guardrails: Do not fabricate results. If no items, suggest query refinement.

3. news_briefs - Headlines with Summaries

Returns headlines with short AI-generated summaries.

Input Parameters: Same as news_headlines

Additional Output Fields:

When to Use: Retrieve brief summaries alongside headlines for more context without full articles.

4. news_fulltext - Full Articles

Returns complete articles with body text.

Input Parameters: Same as news_headlines

Additional Output Fields:

When to Use: Retrieve full articles for deeper analysis, sentiment analysis, or content extraction.

5. news_fulltext_metadata - Full Articles + Metadata

Returns full articles with enriched metadata for knowledge graphs and compliance.

Input Parameters: Same as news_headlines

Additional Output Fields:

When to Use: Retrieve full articles plus metadata (Wikidata/entity labels, IPTC topics) for knowledge-graph integration or compliance use-cases.

6. Translated Tools

The following tools provide machine-translated outputs. Require additional target_lang_code parameter.

news_headlines_translated

news_briefs_translated

news_fulltext_translated

news_fulltext_metadata_translated

Translation Input:

Translation Output: Each item includes a tr object:

{ "id": "...", "title": "Original Title", "tr": { "title": "Translated Title", "description": "Translated Description", "source_lang_code": "en", "target_lang_code": "ja" } }

Common Patterns

Pattern 1: Lookup Ticker, Then Get News

1. Call resolve_names with name="SoftBank" 2. Extract ticker="9984" from results 3. Call news_headlines with ticker="9984"

Pattern 2: Monitor Latest News

1. Call news_headlines with ticker="9984", sort="latest", count=20 2. Process results 3. If next.search_after exists, call again with search_after cursor

Pattern 3: Multi-language News

1. Call news_headlines_translated with ticker="6758", target_lang_code="ja" 2. Use tr.title for Japanese headline, title for original

Pattern 4: Deep Analysis with Metadata

1. Call news_fulltext_metadata with ticker="7974", count=10 2. Extract entities, wikidata, mediatopics for knowledge graph 3. Analyze body text for sentiment or key topics

Pagination Strategy

When sort=latest:

  1. Initial request returns next.search_after in response
  2. Use search_after value in subsequent requests
  3. Keep sort and count consistent
  4. Stop when next object is absent

Example:

Request 1: ticker=7203, sort=latest, count=20 Response 1: {data: [...], next: {search_after: "cursor1"}} Request 2: ticker=7203, sort=latest, count=20, search_after="cursor1" Response 2: {data: [...], next: {search_after: "cursor2"}} ...continue until next is absent

Sort Modes

sort=latest (default)

sort=relevance

Credits & Usage Tracking

All tools return a credits object:

{ "credits": { "rate": 1, "quantity": 15, "total": 9850, "request_id": "req_abc123" } }

Credits per Tool:

Monitor Usage:

Error Handling

Tools may fail due to:

Recommended Strategy:

  1. Validate ticker with resolve_names first
  2. Start with small count (5-10) for testing
  3. Handle empty data arrays gracefully
  4. Log request_id for support requests

Best Practices

  1. Start with resolve_names - Validate ticker before news queries
  2. Use appropriate depth - Headlines for scans, fulltext for analysis
  3. Filter by language - Use lang_code to reduce noise
  4. Filter by country - Use country_code for regional focus
  5. Paginate wisely - Use search_after with sort=latest for chronological feeds
  6. Monitor credits - Check credits.total regularly
  7. Cache results - Avoid redundant calls for same ticker/time
  8. Handle empty results - Suggest query refinement to user

Example Workflow: Track Toyota News

Step 1: Verify ticker Tool: resolve_names Input: {name: "Toyota"} Output: {items: [{ticker: "7203", name: "Toyota Motor Corporation"}]} Step 2: Get latest headlines Tool: news_headlines Input: {ticker: "7203", lang_code: "en", count: 20, sort: "latest"} Output: {data: [...20 headlines...], credits: {...}, next: {search_after: "..."}} Step 3: Get detailed briefs for top stories Tool: news_briefs Input: {ticker: "7203", lang_code: "en", count: 5, sort: "relevance"} Output: {items: [...5 briefs with summaries...], credits: {...}} Step 4: Deep dive into one article Tool: news_fulltext_metadata Input: {ticker: "7203", count: 1} Output: {items: [{id, title, body, entities, wikidata, mediatopics, ...}], credits: {...}} Step 5: Translate for international audience Tool: news_headlines_translated Input: {ticker: "7203", target_lang_code: "ja", count: 10} Output: {items: [{title: "...", tr: {title: "...", target_lang_code: "ja"}}], credits: {...}}

Integration Tips

For Trading Platforms

For Investment Research

For Multilingual Dashboards

For Alert Systems

Support

Docs: https://www.pressmonitor.com/en/docs/news-xjpx-api-mcp Email: support@pressmonitor.com