# Quantral in Claude and ChatGPT > Quantral's MCP server connects an AI assistant to Quantral's stock sentiment data: ask what retail investors and market commentators are saying about a stock, and get scores and monthly recaps back as interactive cards in the chat, with the mentions behind each score one question away. It also reads Quantral's tracked strategies: what each list holds, what changed in its weekly update, and its record against the S&P 500. Full page: https://quantral.com/mcp · Server URL: https://app.quantral.com/api/mcp (remote MCP over Streamable HTTP) · Setup: https://app.quantral.com/connect Connecting requires signing in with a Quantral account, and agent access is included with every plan (the yearly plan has a 7-day free trial). The connection is read-only: an agent can read signals, scores, recaps, and strategies, and can never change the account, watchlists, or billing. It also learns the account holder's name and email so it knows whose account it is reading, and nothing else. ## Things to ask - "What names are trending in the last 24 hours?" - "What is the signal score for NVDA?" - "Summarize last month's chatter recaps for AMD." - "Which names had the strongest 7-day chatter?" - "Show me the mentions behind Tesla's score." - "What changed in the Capitol strategy this week?" - "How has FinX stars done against the S&P 500?" ## What the agent can read 1. `get_top_signals` (**Top signals**): The stocks with the strongest chatter right now, ranked 0 to 100, over either a 24-hour or a 7-day window. 2. `search_companies` (**Company search**): Look up a company by name or ticker, so the agent knows which one you mean before it pulls anything else. 3. `get_company_score` (**Signal strength**): One company's score for both windows, each with its tier word attached: Exceptional, Strong, Moderate, Weak, or Negative. 4. `get_company_recaps` (**Monthly recaps**): Month-by-month summaries of what the conversation around a company has actually been about. 5. `get_company_signals` (**The mentions behind a score**): The individual posts a score was built from, up to 50 per request, so you can read the case instead of taking the number on trust. Korean broker research notes shared on Telegram appear here too, as context only: they do not feed the score. 6. `list_strategies` (**Strategies**): Quantral's tracked strategies: rule-based 10-stock model portfolios, rebuilt weekly and compared with the S&P 500. Each comes with how often it updates, its live record and hypothetical backtest kept separate, the latest list date, and the next update. 7. `get_strategy` (**Strategy**): One strategy's current list, which names are new this update, how each has moved since it joined, and what was dropped. Key stats keep live and backtest apart, with a plain summary and the caveats. 8. `get_strategy_changes` (**Strategy changes**): What was added and dropped in the latest weekly update, or the net changes since a date you give, for checking in on a strategy you already watch. Every score the agent reads is aggregated social sentiment from the sources Quantral tracks, weighted by source and author credibility. It is not a price prediction, and it is not investment advice. ## Also built for agents that trade An AI trading agent or bot can call the same tools your chat does: pull the ranked signals on a schedule, check the chatter around a stock before its own model acts, and read the mentions behind a score. Sentiment is one input; the trading system decides its weight. - Hosted: A remote server you point your agent at. Nothing to run or deploy. - Auth: OAuth sign-in with a Quantral account. No API keys, read-only access. - Rate limit: 60 requests a minute, more than a scheduled strategy needs. ## Interactive cards in the chat - **Ranked top signals**: The current leaderboard of chatter, scored and ordered for the window you ask about. - **Company score**: One company's signal strength across both windows, with the tier word beside each score. - **Monthly recaps**: The month-by-month story of a company, the same recaps you get on its page in the app. - **Strategies**: A strategy's list, what its latest weekly update added and dropped, and its record against the S&P 500, live and backtest kept apart. ## Setup Works with any MCP-compatible agent. ### Claude 1. Open Settings, then Connectors. 2. Choose Add custom connector. 3. Paste the server URL. 4. Sign in with your Quantral account and approve. Or skip the URL and add Quantral from [Claude's connector directory](https://claude.ai/directory/691982a8-b3cf-481d-b630-abd15a1af6b0). ### ChatGPT 1. Open Settings, then Apps and connectors, then Advanced, and enable developer mode. 2. Add a connector with the same server URL. 3. Sign in and approve. ### Or let the agent connect itself Paste this into the chat: > Connect the Quantral MCP server. Endpoint: https://app.quantral.com/api/mcp (remote MCP over Streamable HTTP, OAuth sign-in with my Quantral account). If you can add MCP servers or custom connectors yourself, add it now and walk me through the sign-in. If you cannot, list the exact steps to add it in this app, then use its tools to tell me which names have the strongest chatter right now. ## Frequently asked questions ### What is an MCP server? MCP, the Model Context Protocol, is an open standard that lets an AI assistant connect to an outside service and read data from it. Quantral's MCP server is the endpoint your assistant connects to, so Claude, ChatGPT, or any MCP-compatible agent can pull signal scores, the mentions behind them, monthly recaps, and Quantral's tracked strategies straight into the conversation. ### Can Claude or ChatGPT analyze stocks with live data? Connected to Quantral, yes: your assistant reads the live signal score for a stock, the individual posts behind it, and its monthly recaps, and reasons over them in the conversation. The analysis stays grounded in sentiment data: what the crowd is saying, weighted by credibility, not a prediction of where the price goes. ### Is it included in my subscription? Yes. Agent access is included with every Quantral plan, weekly, monthly or yearly, at no extra cost. You connect by signing in with your Quantral account, and an active subscription (or your 7-day free trial on yearly) is what keeps the connection working. ### Can the agent change anything in my account? No. The connection is read-only. An agent can read signals, scores, recaps, and strategies, and it can never change your account, your watchlists, or your billing. It also learns your name and email so it knows whose account it is reading, and nothing else. ### Is this investment advice? No. Every score the agent reads is aggregated social sentiment from the sources Quantral tracks, weighted by source and author credibility. It describes how loud and how credible the conversation is right now, not where a price is going, and nothing your assistant reports back is a recommendation to buy or sell. ### How current is the data? The server queries Quantral live, so your assistant reads the current scores for both the 24-hour and the 7-day window every time you ask, along with the mentions sitting behind them. ### Are there any limits? The server accepts up to 60 requests a minute per account, which is well beyond what a normal conversation uses. Pulling the individual mentions behind a score returns up to 50 of them per request. --- Public data, scored. Not financial advice, do your own research.