Alpha Leak is a production-grade, multi-phase trading intelligence system built for Pump.fun on Solana. It processes every on-chain event as it lands, runs layered analysis across wallet scoring, adversarial detection, ML inference, and market regime classification, then executes trades atomically when high-conviction signals emerge.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/0xW1re/solvedocs/llms.txt
Use this file to discover all available pages before exploring further.
Introduction
Understand what Alpha Leak does, the problems it solves, and the system scale at steady state.
Architecture
Full pipeline map — from raw on-chain events to live execution across 30+ concurrent services.
Quickstart
Connect to the live signal feed and query historical data in under five minutes.
API Reference
Complete reference for all REST endpoints and the live SSE stream.
How the pipeline works
Alpha Leak operates as a single process hosting 30+ concurrent background services, organized into four named phases that build on each other.Ingest on-chain events
The pipeline receives every trade, token creation, and graduation event the moment it lands on-chain. No polling, no batching — the
EventProcessor decodes and fans out each event to downstream services immediately.Score wallets and detect threats
Phase 1 builds 7-component alpha scores for every wallet observed. Phase 2 runs intelligence services — bundle detection, creator risk scoring, co-occurrence graphs, and lifecycle classification — to establish the adversarial landscape around each token.
Run ML inference
Phase 3 scores every signal through LightGBM models compiled to ONNX, running in-process at sub-millisecond latency. Models are calibrated with Platt scaling so their probability outputs are directly actionable as strategy thresholds.
Explore by section
The Pipeline
Deep dives into ingestion, Phase 1 scoring, Phase 2 intelligence, and Phase 3 ML inference.
Intelligence
Wallet profiling, adversarial detection, genesis scoring, and market regime classification.
ML System
ONNX model deployment, the 68-feature vector reference, and training methodology.
Live Trader
Strategy configuration, phase gating, circuit breakers, and performance monitoring.