Alphai Mcp — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Alphai Mcp (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Real-time, AI-enriched financial news for AI agents and trading bots — over the Model Context Protocol. Hosted at `mcp.alphai.io`, no install, OAuth (no API key to paste), free tier 100 calls/day.
Every story is enriched with per-ticker analysis, a category (14 buckets), and a 1–10 relevance score, so an agent can filter to what actually matters before spending a reasoning token.
This repo is the public home +server.jsonmanifest of the hosted AlphaAI MCP server (the listing on Smithery, Glama, mcp.so, the MCP Registry and Awesome MCP Servers). The product is AlphaAI — a financial-news platform built for AI agents. There is nothing to self-host: to use it, just connect tohttps://mcp.alphai.io/mcp.
The server speaks Streamable HTTP at https://mcp.alphai.io/mcp. Add it to any OAuth-capable MCP client. Claude Code:
claude mcp add --transport http alphai https://mcp.alphai.io/mcpThe first tool call opens a browser for OAuth 2.1 (DCR + PKCE) — a login, no key to copy-paste. ChatGPT, Claude Desktop / claude.ai, Cursor, VS Code, Windsurf and Gemini connect the same way. Or use the one-click listing on Smithery.
| Client | Config |
|---|---|
| Claude Desktop / claude.ai | Settings → Connectors → Add custom connector → https://mcp.alphai.io/mcp |
| Cursor | ~/.cursor/mcp.json → { "mcpServers": { "alphai": { "type": "http", "url": "https://mcp.alphai.io/mcp" } } } |
| VS Code Copilot | .vscode/mcp.json → { "servers": { "alphai": { "type": "http", "url": "https://mcp.alphai.io/mcp" } } } |
| Generic | Streamable HTTP, URL https://mcp.alphai.io/mcp, OAuth 2.1 |
| Tool | What it does |
|---|---|
alphai_news_search | Full-text + filtered news search (query, tickers, category, dates, relevance) |
alphai_ticker_news | Latest news for one ticker (optionally incl. insider) |
alphai_trending | Biggest stories of the last 48h by relevance |
alphai_actionable_now | Breaking, decision-grade news (actionability + novelty gate) |
alphai_insider_news | SEC Form 4 insider trades + 13F ownership moves as news |
alphai_pair_analysis | Two-ticker read-across (news naming both companies) |
alphai_article | Fetch a single article by uid |
alphai_tickers | Discover supported tickers — US stocks, ETFs, crypto & foreign listings |
alphai_alerts_list / _subscribe / _unsubscribe | Manage your own ticker alert subscriptions (Basic/Pro) |
All tools are read-only except the alphai_alerts_* writes, which only ever touch the caller's own subscriptions. Full schemas, params and defaults are advertised by the server (annotations included) and documented at alphai.io/mcp.
| Free | Basic | Pro | |
|---|---|---|---|
| Price | $0 (no card) | $2.99/mo | $9.99/mo |
| Rate limit — burst | 20 / min | 60 / min | 300 / min |
| Rate limit — daily | 100 / day | 10,000 / day | 100,000 / day |
| Alert tools | — | ✓ | ✓ |
| Page size | 10 | 10 | up to 50 (bulk) |
OAuth 2.1 with Dynamic Client Registration (RFC 7591) and PKCE per the MCP authorization spec. Compatible clients discover the OAuth metadata automatically — no manual API key setup. Tool discovery (tools/list) is open; calling a tool requires the OAuth login.
Drop-in skills that drive these tools (stock brief, market pulse, insider radar, peer read-across, manage alerts): [makeev/alphai-claude-skills](https://github.com/makeev/alphai-claude-skills).
io.github.makeev/alphai-mcphttps://mcp.alphai.io/mcp; thereis nothing to self-host. This repo is the catalog home + server.json manifest.
raw_text (full article bodies) is never served — copyright. Responses carry titles,AI summaries, per-ticker analysis, categories and relevance scores.
MIT licensed. Built by AlphaAI.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.