MT Newswires

By MT Newswires

Get New App Alerts

Open in ChatGPT

Open in Claude

Overview

MT Newswires brings real-time global financial news directly into ChatGPT — providing original, unbiased and comprehensive multi-asset class coverage of capital markets and economies in over 180 topics. Powered by a global newsroom trusted by the largest banks, brokerages, professional trading, wealth management and research applications globally, every story is built for AI — so you get answers grounded in live market intelligence, not stale data. What you can do: • Get the latest news on any ticker and analyze price impact • Monitor energy sector developments across global equities • Track macro events and global central bank signals in real time • Screen for earnings, M&A, and market-moving headlines Example prompts: → "What is the latest news for NVIDIA and has it influenced the stock's price?" → "Summarize the major analyst rating change developments from the past hour." → "What macro events are driving markets today?"

Tools

fetch_news

ChatGPT

Fetch a single MT Newswires news article by its transmission ID. Returns the full article content including headline, summary, symbols, source, and metadata.

fetch_news_batch

ChatGPT

Fetch multiple MT Newswires news articles by their transmission IDs in a single call. More efficient than fetching one at a time. Returns full article content including headline, summary, symbols, source, and metadata for each article.

fetch_news_by_security

ChatGPT

Fetch MT Newswires news articles for a specific company or ticker symbol, sorted newest to oldest. Use this tool INSTEAD OF search_news when the user asks for news about a specific company, ticker, or security (e.g. 'news for AAPL', 'latest Apple news', 'what's happening with Tesla'). Only use search_news when the user asks a topical question not tied to a specific security (e.g. 'AI regulation news', 'interest rate decisions'). Returns headlines, summaries, and transmission IDs. You MUST call format_news before presenting results to the user.

format_news

ChatGPT

Format news articles for presentation to the user. This tool MUST be called as the final step after using search_news, fetch_news, or fetch_news_by_security. Never present news results to the user without calling this tool first. Present the formatted_text output directly to the user. Source citation is MANDATORY — every article must include a citation in the format: (Source: MT Newswires <transmission_id>). When presenting multiple articles: group by date (newest first), deduplicate articles covering the same event, and use clear headers and bullet points. Presentation quality matters — make it clear and visually pleasing.

list_metacodes

ChatGPT

Discover which MT Newswires financial-news categories are available for a search news request — sectors (energy, oil & gas, natural gas, technology, healthcare, financials, utilities, industrials), commodities and asset classes (crypto, bonds, forex, ETFs, options), regions and countries, the Fed / interest rates / monetary policy, and corporate events (M&A, IPOs, earnings, dividends, buybacks, analyst upgrades/downgrades, FDA decisions). Returns the catalog of MT 'metacode' classification codes, each with a title and description, grouped by category. Call this EARLY — before search_news or search_news_by_metacode — for ANY news request that could map to a category, to get the exact codes. Only call it once per conversation; if you already have the codes, reuse them. Codes are exact and include the .MN suffix (e.g. NAT.MN, FEDN.MN, IPON.MN). The output also includes a 'combos' list — ready-made code bundles for common topics (crypto, energy, the Fed, M&A, earnings, etc.); when a request matches one, pass its codes together to search_news_by_metacode.

onboarding

ChatGPT

Return a guided onboarding walkthrough — an intro, capability sections (each with a one-line lead and a couple of example queries), practical phrasing tips, and a closing invite. Call this tool when the user asks what you can do, for help, how to use this, for examples, or when they seem unsure how to phrase a financial-news question. Present it to the user as a friendly walkthrough: render each section as a markdown header with its lead text, the example queries as block-quoted copy-pasteable prompts beneath, and the tips as a bulleted list. Keep the tone warm and inviting, not encyclopedic.

search_news

ChatGPT

Search MT Newswires financial news articles using semantic search for topical queries (e.g. 'AI regulation impact', 'interest rate decisions', 'supply chain disruptions'). Do NOT use this when the user asks for news about a specific company or ticker — use fetch_news_by_security instead. Returns headlines, dates, and relevance scores. After fetching articles with fetch_news, you MUST call format_news before presenting any results to the user.

search_news_by_metacode

ChatGPT

Search MT Newswires financial news by category for the latest articles in a topic area — sectors (energy, oil & gas, natural gas, technology, healthcare, financials), commodities and asset classes (crypto, bonds, forex, ETFs, options), regions and countries, the Fed / interest rates / monetary policy, and corporate events (M&A, IPOs, earnings, dividends, buybacks, analyst upgrades/downgrades, FDA decisions, trading halts). Retrieves recent articles (newest first) tagged with one or more MT 'metacode' codes; multiple codes are combined as a union by default (use match='all' for intersection). Provide an optional 'query' to semantically rank and filter results WITHIN the selected categories — the way to target a sector or topic that has no metacode of its own (e.g. metacodes=['MERGA.MN'], query='biotech' for biotech M&A; metacodes=['MERGA.MN'], query='healthcare' for healthcare M&A). Without a query, results are the most recent in the category. ALWAYS use this in coordination with search_news — run BOTH and combine: this gives the latest in the category, search_news gives the most semantically relevant. For a named company or ticker use fetch_news_by_security. To narrow to where two categories overlap (e.g. healthcare analyst upgrades), set match='all' to intersect the codes (only articles tagged with EVERY code); match defaults to 'any' (union). Intersection works with or without a query — with a query it ranks by relevance, without one it returns the newest in the overlap. IMPORTANT: metacode tagging is imperfect — articles are sometimes mistagged. Review each result against the user's original request and silently drop any that are not genuinely relevant; present only the articles that actually relate to what the user asked for. Unless you have already called list_metacodes earlier in this conversation, call it FIRST to load the exact codes (only once per conversation; reuse the codes if you already have them). Codes must be exact and include the .MN suffix (e.g. NAT.MN, FEDN.MN, IPON.MN, MERGA.MN, ANACU.MN, CRYP.MN). Call list_metacodes for the full catalog. You MUST call format_news before presenting any results to the user.