Confidence over hype
A five-factor formula scores every call 0–100. Below the threshold, the No Signal Protocol forces the result to neutral — no weak signal ever leaks as a buy or sell.
confidence: 41 → neutral (forced, < 50)
A Model Context Protocol server that runs nine theory-based technical-analysis engines on demand. Every result ships with a 0–100 confidence score, an explicit invalidation level, and named author attribution — so your AI agent can act on the call or know exactly when to bail.
A five-factor formula scores every call 0–100. Below the threshold, the No Signal Protocol forces the result to neutral — no weak signal ever leaks as a buy or sell.
Every positive signal returns the exact price condition that would falsify it. Your agent doesn't have to guess where to set the stop — the engine already did.
Most TA libraries quietly use future bars in historical classification. We don't. The public TA audit documents the fix to VSA's leakage, with reproducing tests.
Each engine returns the standard AnalysisResult contract.
Theory names are attributed: Dow, Hosoda, Wyckoff, Elliott, Chan Zhongshu,
Scott Carney, J. Peter Steidlmayer.
Scanners filter the universe; memecoin_discovery
surfaces trending DEX tokens across 200+ networks. Hand the address
straight to any analyzer.
{
"mcpServers": {
"sigmapilot": {
"command": "uv",
"args": [
"tool", "run", "--from",
"git+https://github.com/SigmaPilotAI/sigmapilot-mcp.git",
"sigmapilot-mcp"
]
}
}
}
# Claude.ai → Settings → Connectors URL: https://<your-deployment>/mcp Transport: streamable-http Auth: OAuth2 / JWT via Auth0 # Deploy your own in minutes via Railway # https://github.com/SigmaPilotAI/sigmapilot-mcp#deploy