Purpose
AlphAI MCP is a Model Context Protocol (MCP) server designed for AI agents and trading bots to retrieve financial news, stock market sentiment, and SEC Form 4 insider trading data. Its core idea is that each news item, before being returned to the agent, has already been augmented with per-stock analysis, 14 categories of classification tags, and a relevance score from 1-10, enabling the agent to filter first and then read, thereby saving inference costs.
Main Capabilities
The server provides a set of read-only tools (except for alert subscription management), covering the following areas:
- News search and retrieval: alphai_news_search supports full-text and filtered search (query terms, ticker symbols, categories, dates, relevance); alphai_ticker_news retrieves the latest news for a single ticker symbol (optionally including insider trading); alphai_trending returns trending articles from the past 48 hours ranked by relevance, with republished content collapsed.
- Story aggregation: alphai_stories aggregates events reported by two or more publishers; list=top ranks the past 48 hours, list=latest paginates with newest first; alphai_story retrieves a single story by story_id with its summary, sourced key facts, per-stock interpretation, open questions, and the underlying articles.
- Decision-grade news: alphai_actionable_now provides breaking, actionable news (filtered by actionability and novelty thresholds).
- Insider trading: alphai_insider_news provides SEC Form 4 insider trading news; alphai_insider_clusters identifies issuers in which multiple insiders bought in a single round of open-market purchases (labeled by buyer).
- Analysis and earnings: alphai_pair_analysis provides cross-interpretation of two ticker symbols (news mentioning both companies); alphai_article retrieves a single article by uid (adding structured earnings interpretation for SEC filings); alphai_earnings provides AlphAI's filing-verified per-stock earnings interpretation and next report date.
- Macro and calendar: alphai_calendar provides scheduled macro releases (CPI, FOMC, unemployment claims) with their timing and follow-up coverage; alphai_macro provides macroeconomic developments (data, central banks, interest rates, forex, commodities).
- Ticker symbols and monitoring: alphai_tickers discovers supported ticker symbols (US stocks, ETFs, cryptocurrencies, and foreign listings, each with its own next_report_date); alphai_watchlist_brief provides recent news and filing events for saved watchlists (available on all plans); alphai_radar provides news activity and tone for the whole market or a watchlist, with evidence links, 4-hour/24-hour windows, and snapshot-pinned pagination.
- Research retrieval: search / fetch provide citable article discovery and retrieval for compatible research clients.
- Alert management: alphai_alerts_list / alphai_alerts_subscribe / alphai_alerts_unsubscribe manage the user's own ticker alert subscriptions (Basic/Pro plans).
How to Use
The server is hosted at mcp.alphai.io and requires no self-deployment. It uses Streamable HTTP, at the address https://mcp.alphai.io/mcp。将其添加到任何支持 any MCP client with OAuth.
Taking Claude Code as an example, connecting will open a browser for OAuth 2.1 authentication (DCR + PKCE); you only need to log in, with no need to copy and paste keys. ChatGPT, Claude Desktop / claude.ai, Cursor, VS Code, Windsurf, and Gemini connect in the same way. A one-click listing on Smithery is also available.
For JSON configuration clients (such as Cline, Cursor, etc.), configure the MCP server URL as:https://mcp.alphai.io/mcp
OAuth 2.1 supports Dynamic Client Registration (RFC 7591) and PKCE, compliant with the MCP authorization specification. Compatible clients automatically discover OAuth metadata, with no manual API key setup required. Tool discovery (tools/list) and invocation both require authentication. Headless clients can use the same API key as the REST API in the Authorization: Bearer ak_live_… header; never put keys in URLs. OAuth and API keys have tool parity.
Additionally, there are ready-made Claude Code skills (Stock Brief, Market Pulse, Insider Radar, Peer Cross-Read, Management Alerts) at makeev/alphai-claude-skills.
When to Use
- Trading bots and AI agents need to filter financial news and process only highly relevant content.
- Need to monitor news, insider transactions, or earnings information for specific stock tickers.
- Need to track the macroeconomic event calendar or overall market sentiment.
- Need to manage watchlists and receive related alerts.
Limitations
- The free tier is limited to 20 requests per minute and 100 requests per day.
- Radar refreshes every minute; latency applies to the full snapshot (including supporting headlines); collection and enrichment add processing time. Its descriptive scores are not predictive or confirmed alerts. Radar accepts up to 100 rows per call (all plans), with no separate stock ticker quota; saved watchlists retain their original limits (10 companies on free, unlimited on paid).
- raw_text (full article body) is not provided, for copyright reasons; responses include only headlines, AI summaries, per-stock analysis, categories, and relevance scores.
- The tools provide news summaries, not buy/sell recommendations.
- All tools are read-only, except for alphai_alerts_* write operations, and these affect only the caller's own subscriptions.
Additional Information
- Full schemas, parameters, and defaults are published by the server (with annotations) and documented at alphai.io/mcp.
- Playground and documentation:https://alphai.io/mcp
- REST API and SDKs (Python + TypeScript):https://alphai.io/developers
- Smithery listing:https://smithery.ai/servers/mihail-makeev/alphai-news
- Glama connector:https://glama.ai/mcp/connectors/io.github.makeev/alphai-mcp
- MCP Registry: io.github.makeev/alphai-mcp
- Changelog:https://alphai.io/changelog
- Licensed under MIT, built by AlphAI.
