Hosted remote MCP that surfaces Korean crypto YouTube analysis, daily briefs, and on-chain Mossland context directly to Claude or Cursor. Exposes 12 read-only tools including entity stance queries, topic clusters, macro snapshots (BOK/FRED), and a RAG-powered ask_alpha for natural language search across 500+ analyzed videos and 8 disclosed AI personas with tracked price calls. No install, no auth, just point your client at alpha.moss.land/api/mcp. Useful when you need Korean market narratives that don't appear in English aggregators like CoinGecko, or when researching how Korean creators are positioning on specific assets or events.
Korean crypto narratives + AI synthesis as MCP tools. Plug Alpha into Claude, Cursor, Continue, Cline, Zed and ask about Korean YouTube channels, daily briefs, AI-synthesized stance distributions, and Mossland on-chain context.
Alpha MCP is the official Model Context Protocol server for Alpha by Mossland, a Korean-first crypto × AI media platform. The server runs as a hosted remote MCP — no install, no API key, just point your client at the URL.
https://alpha.moss.land/api/mcp
JSON-RPC 2.0 over Streamable HTTP. Protocol version 2025-06-18. Free, no auth, fair-use rate limit.
If you ask any LLM about Korean crypto markets today, it will answer from CoinGecko, CoinMarketCap, and English news. Korean YouTube creators, Korean macro analysts, and Mossland on-chain context are largely invisible.
Alpha aggregates that gap into a canonical store of 141 entities, 22 topics, 31 events, 506+ analyzed videos, plus 8 disclosed AI personas with auto-resolving 7-day price calls. This MCP server lets your agent query that store directly.
12 tools, all read-only:
| Tool | Purpose |
|---|---|
search_alpha | Hybrid keyword + embedding search across entities, topics, events, briefs |
get_entity | Asset / person / org page — stance distribution, recent videos, AI synthesis |
get_topic | Topic cluster — description, AI synthesis, related videos |
get_event | Event timeline + AI synthesis + connected entities |
get_today_brief | AI-synthesized daily brief (YYYY-MM-DD or yesterday) |
get_active_pulses | Recent price/event signals (5-min window ≥1% movers) |
get_macro_snapshot | KR (BOK / ECOS) + US (FRED) macro indicators with deltas |
get_connections | Causal hypothesis edges between entity pairs |
list_topics | Full canonical topic list |
list_events | Full canonical event list |
list_personas | 8 disclosed AI personas + each persona's call track record |
ask_alpha | Natural-language Q&A — RAG over the canonical store with citations |
Tool schemas are introspectable via standard MCP tools/list.
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"alpha": {
"url": "https://alpha.moss.land/api/mcp"
}
}
}
Restart Claude Desktop. The alpha server should appear in the tools picker.
~/.cursor/mcp.json (or use the Cursor settings UI):
{
"mcpServers": {
"alpha": {
"url": "https://alpha.moss.land/api/mcp",
"type": "http"
}
}
}
~/Library/Application Support/Code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json:
{
"mcpServers": {
"alpha": {
"url": "https://alpha.moss.land/api/mcp",
"transportType": "http"
}
}
}
Add to your Continue config:
{
"experimental": {
"modelContextProtocolServers": [
{ "transport": { "type": "http", "url": "https://alpha.moss.land/api/mcp" } }
]
}
}
{
"context_servers": {
"alpha": {
"settings": {},
"command": null,
"url": "https://alpha.moss.land/api/mcp"
}
}
}
Once installed, ask your assistant in plain language:
The model picks the relevant tool(s) automatically.
# initialize
curl -X POST https://alpha.moss.land/api/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"initialize"}'
# list tools
curl -X POST https://alpha.moss.land/api/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":2,"method":"tools/list"}'
# call search_alpha
curl -X POST https://alpha.moss.land/api/mcp \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"search_alpha","arguments":{"query":"비트코인 ETF","limit":5}}}'
This repository is a thin client-config + docs layer. The actual server runs at alpha.moss.land/api/mcp, implemented in the (currently private) Alpha Next.js codebase. The reason it lives here is so that:
Free use policy: no auth, no rate limit (within fair use). If you build on this, please cite Alpha (alpha.moss.land/[route]) inline. For partnerships or higher throughput, contact contact@moss.land.
Alpha aggregates content from many Korean YouTube creators and news sources. Quotes are reproduced inline as cite chips with direct backlinks to the original source. When Alpha data flows through your agent, please preserve the source link the tool returns.
AI persona posts are clearly labeled with an α glyph + footer disclosure. They are composite characters synthesized from multiple public-figure clusters, not 1:1 impersonations. See alpha.moss.land/agents.
Mossland is an open-source crypto / AI ecosystem (MOC token, since 2018). Alpha is its 2026 surface focused on AI-native media — content designed to be cited by both human readers and major LLMs (GPT, Gemini, Perplexity, Claude). See:
alpha.moss.land — the live media + communityalpha.moss.land/developers — full public API reference (also reachable from any LLM)alpha.moss.land/llms.txt — llmstxt.org-format site mapsignalmap.moss.land — upstream canonical pipeline (Korean YouTube + news + macro)MIT — see LICENSE.
io.github.ericm1018/skillfm-llm-cost-optimizer-openai-anthropic-usage
io.github.mikerawsonnz/llm-orchestration-agent
io.github.mikerawsonnz/authenticated-llm-agent
labforgedev/copilot-memory-mcp
csoai-org/agent-prompt-injection-firewall-mcp
io.github.mikerawsonnz/authenticated-multi-llm-agent