axe-fx-v3 — Solana Memecoin Research Stack
Disclaimer: strictly private research and learning project. Not investment advice, not financial advice, not a commercial offering. No reliable alpha found so far.
Use case
My own Solana memecoin hypotheses are captured by four discovery collectors, run through a 16-gate pipeline, persisted in 23 axe_v3.* tables, and evaluated against strategy functions. The paper trader runs continuously; the actual executor stays live-locked behind a kill switch — live activation only after 60+ days of data, multi-regime validation, and a manual key flip.
Lineage
- solscout — first explorations in Solana memecoin territory. Discovery only, no trading.
- axe-fx-platform v1 — first solid setup with a real order lifecycle. Ran stably as a paper trader and was the learning platform for the pain points v3 now addresses charter-first.
- axe-fx-platform v2 — optimization iteration together with Axiom Extension v2. Never fully shipped to production, but produced the insights that drove the v3 rewrite.
- axe-fx-v3 — current charter-first rebuild. v2 has been cleanly decommissioned since Phase 8, data migrated into
axe_v3.
What's done so far (v3)
- Phases 0–8 complete (foundation, discovery, persistence, infrastructure, pipeline, UI/API, strategy, trading, decommission)
- 4 discovery collectors, 6 service containers, 16 pipeline gates, 23 DB tables
- Paper trader live since Phase 7
- v2 cleanly decommissioned, data closure into
axe_v3 - Strategy lab with champion search (
telegram+early+sl_5) — train positive, test negative → strategy rejected, no blind go-live - Executor safety defaults:
kill_switch=1,enabled=0,max_stake_sol=0.05forced
In progress
Longer observation windows for the data pipeline, multi-regime validation, honest backtest-live parity without cherry-picking, strategy lab iterations until either a reliable alpha emerges — or the hypothesis is cleanly disproven.