Open quant research that shows its work.
qMachina is an active open-source quant engineering desk, not a static signal feed. Founded by AI engineers from Apple, Intel and IBM, we take on the market-data infrastructure, C++ replay, AI research automation and daily adversarial review so the community can propose ideas, inspect evidence and see which hypotheses survive. Every signal family is continuously evaluated and gated by its own recent performance; every call is timestamped; every failure is published.
Built to open what closed quant shops hide: infrastructure, evidence, failures and process. Verify us — don't trust us.
Studies — the research, drawn at tick resolution
Open the actual interactive charts behind the desk. Each is rendered straight from our C++ Level-3 real-fill replay — independently re-derivable, no cherry-picked screenshots.
Open ↗
Regime & flip machine
What we see: 5s / 5m / 15m regime ribbons under the candles, with every state flip logged. Click any flip to read the exact math that fired it.
Interactive · NQ 04-27
Open ↗
5s VWMA ribbon
What we see: a six-band volume-weighted moving-average ribbon (fast→slow) on 5-second traded candles — the timing layer the flip machine reads for momentum.
Interactive · NQ 04-27
Open ↗
Entry-signal overlay
What we see: a 250 ms micro-momentum entry signal painted on the ribbon — every long/short trigger marked, with its causal evidence in the table below.
Interactive · NQ 04-27
Open ↗
HD give-back charts
What we see: full-session 1-second candle charts at 23,562px wide — every trade, SL/TP and exit reason marked, across 81 train + holdout days of the give-back exit study.
HD SVG galleryThumbnails are real frames of the live charts. All performance is SIMULATED — see the Rule 4.41 disclosure below.
The record, drawn from the data
Every chart on this page is generated server-side from the actual research and live artifacts — no cherry-picked screenshots. All performance is SIMULATED; see the Rule 4.41 disclosure below.
The qMachina story
A living AI engineering desk
Members are joining an active research team, not downloading a finished PDF. The book evolves every day as new hypotheses, market observations and failure cases enter the queue, then get converted into tests, artifacts and public decisions.
We carry the infrastructure burden
Databento MBO ingestion, session caches, C++ Level-3 real-fill replay, dashboards, Discord alerts, JSONL feeds and verification tools are the heavy machinery. The point is to make hedge-fund-grade research infrastructure inspectable and usable by an open-source community.
Every family, continuously evaluated
ORB flips/fades, retest failures, prior-day & overnight level crosses — each detected at tick resolution and scored against its own out-of-sample history. We don't run a setup because it once worked; we run it because it's working now.
Live regime gates
A walk-forward gate benches a family when its trailing-K-session performance turns cold — before the cold streak, not after. The dashboard shows you which of our own signals are ARMED and which are COLD. We sit out, on the record.
Cryptographically verifiable
Every signal is hash-chained at emission time — frozen before the outcome is known. A 30-line verifier anyone can run re-hashes the whole log and proves nothing was altered or back-dated. No screenshots. A chain you can check.
The falsification library
We publish what we killed: latency-alone, market-maker pull on the detector, ORB-gating of confluence, the gap-fill magnet, afternoon ORB, the bull-flag premium. Showing the failures is unheard of in this market. It's also the whole point.
The differentiator: open quant infrastructure, not a black-box feed. Almost every futures-signal service dies on track-record credibility and cold-streak churn. Our regime gate is the built-in answer to "why did you go quiet this week?" — because the family was cold, and we told you and stopped, on the record, in advance. Where closed shops hide the machinery and retail products sell cherry-picked screenshots, qMachina publishes the record, the failures, the research loop and the discipline to sit out.
Every marker is a real entry→exit recorded live today in the shadow log (green win / red loss), positioned by clock time. Net is the day's running result — SIMULATED / paper shadow, not actual fills. The hash-chained record is one click away.
See every signal, hash-chained the instant it fired → the verifiable track record.
Pricing
- Free TradingView Levels indicator — OR, PDH/PDL/PDC and overnight levels
- Morning "what's armed / what's cold" regime post
- ~1 delayed event card/day
- Weekly scorecard, chain link and falsification notes
- Real-time Discord alerts — ORB flip/fade, retest:fail, PD/ON-level fades
- Invite-only TradingView Signals chart aid — bar-close confirmation, not the source of truth
- Entry zone, level, time-box exit
- Per-event PNG chart card and chain-linked record
- Everything in Signals
- Live Regime Dashboard: per-family × instrument gate states (HOT/COLD + trailing-K PnL)
- Mechanics studies: volume-weighted registers, rolling z-scores, VPOC, momentum vs absorption
- Member research room: propose hypotheses, follow build/test/RIGOR loops and see why we sit out
- Multi-instrument full set
- Raw event JSONL/API feed, option-MATH artifacts and GPU/order-flow viewers
- Monthly research review: infrastructure, regimes and candidate ideas
B2B / API event-feed license: custom — contact.
Required Disclosures
CFTC RULE 4.41 — SIMULATED PERFORMANCE
HYPOTHETICAL OR SIMULATED PERFORMANCE RESULTS HAVE CERTAIN LIMITATIONS. UNLIKE AN ACTUAL PERFORMANCE RECORD, SIMULATED RESULTS DO NOT REPRESENT ACTUAL TRADING. ALSO, SINCE THE TRADES HAVE NOT BEEN EXECUTED, THE RESULTS MAY HAVE UNDER-OR-OVER COMPENSATED FOR THE IMPACT, IF ANY, OF CERTAIN MARKET FACTORS, SUCH AS LACK OF LIQUIDITY. SIMULATED TRADING PROGRAMS IN GENERAL ARE ALSO SUBJECT TO THE FACT THAT THEY ARE DESIGNED WITH THE BENEFIT OF HINDSIGHT. NO REPRESENTATION IS BEING MADE THAT ANY ACCOUNT WILL OR IS LIKELY TO ACHIEVE PROFIT OR LOSSES SIMILAR TO THOSE SHOWN.
Trading futures involves substantial risk of loss and is not suitable for all investors. Past or hypothetical performance is not indicative of future results.
qMachina publishes impersonal, broadcast signals identical to all subscribers. It is general information, not individualized trading advice; we do not direct, access, or trade any subscriber account. Our current published track record is paper/shadow (simulated), not a live-fill record.
Read the complete required disclosures — Rule 4.41, full futures & options risk, the impersonal-publisher statement, and data/simulation limitations — on the Disclosures & Risk Disclosure page.