GSS MCP Server

Model Context Protocol server exposing Grey Swan Signals market stress intelligence as callable tools for AI agents.

GSS scores 30+ financial signals across 6 risk vectors to produce a composite market stress index (0-100). Data updates weekly. This server provides read-only access to pre-computed scored outputs — no raw data, no methodology.

Tools

Related MCP server: SignalFuse MCP Server

Risk Vectors

Signal names shown are current as of the live catalog (get_signal_catalog) — call that tool for the authoritative, up-to-date list rather than treating this table as a static reference.

Alert Events

The alert system surfaces what changed — often more valuable than static scores. "Elevated" below always means a signal scoring 51+ (ALERT or CRITICAL):

• STATUS_CROSSING: Signal transitions upward into ALERT or CRITICAL tier

• FAST_MOVER: Signal score rises 10+ points in 7 days

• VECTOR_SURGE: 2+ signals in same vector elevated

• CROSS_VECTOR_CONTAGION: 3+ vectors with at least one signal elevated

• FAST_MOVER_CLUSTER: 3+ fast movers across any vectors in 7 days

• TRANSMISSION_CHAIN: Volatility + Credit Risks both elevated

• SEVERITY_MULTIPLIER: Valuation + any transmission vector elevated

Authentication

Include a Bearer token in the Authorization header:

Authorization: Bearer gss_your_api_key_here

Request a key at greyswansignals@gmail.com with your intended use case.

Tiers

Connection

Claude Desktop

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "gss": {
      "url": "https://mcp.greyswansignals.com/sse",
      "headers": {"Authorization": "Bearer gss_your_key_here"}
    }
  }
}

Anthropic API (Remote MCP)

import anthropic

client = anthropic.Anthropic()

response = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=1024,
    mcp_servers=[{
        "type": "url",
        "url": "https://mcp.greyswansignals.com/sse",
        "name": "gss",
        "authorization_token": "gss_your_key_here"
    }],
    messages=[{
        "role": "user",
        "content": "What is the current market stress reading and which vectors are elevated?"
    }]
)

Data Freshness

Every response includes a meta object:

{
  "data": { ... },
  "meta": {
    "last_updated": "2026-04-06",
    "update_cadence": "weekly",
    "source": "Grey Swan Signals — greyswansignals.com"
  }
}

GSS data updates weekly. For intraday applications, GSS provides macro regime context, not real-time signals.

Exception: get_weekly_narrative fetches the live site on every call rather than a weekly export, so its content can change between calls independent of the meta.update_cadence value on the response (which describes scoring data, not the narrative).

Scoring

Every signal and vector resolves to a 0-100 Signal Score and a status tier: BASELINE (0-25), WATCH (26-50), ALERT (51-75), CRITICAL (76-100). Scoring methodology is proprietary and not exposed via this API.

Self-Hosting

# Install dependencies
pip install -r requirements.txt

# Initialize API keys database
python manage_keys.py init
python manage_keys.py create --label "local-test" --tier free

# Configure
cp .env.example .env
# Edit .env with paths to your GSS data

# Start server
python gss_mcp_server.py
# → INFO: Uvicorn running on http://0.0.0.0:8765

# Test
curl http://localhost:8765/health

Health Check

GET /health

Returns server status, database availability, and last data update timestamp.

Grey Swan Signals — greyswansignals.com