Product
18 engines and an analyst that explains them.
Quantitative engines measure risk, factors, regimes and attribution. The AI Analyst explains the results in plain language, as analysis of observed data, not guesses.
- Read-only API keys
- Never places trades
- Free tier at launch
Architecture
Data in. Research and answers out.
Connected with read-only keys. Amstos never places orders, withdraws funds or holds custody.
- 01In build
Data
Market data, portfolios, transactions and trades in one view.
- 02In build
Quantitative engines
Portfolio, risk, factor, regime, attribution and strategy metrics. Mostly statistical and simulation models, not black boxes.
- 03In build
Research & backtesting
Test strategies, sweep parameters and check robustness.
- 04In build
AI Analyst
Reads the engine results and explains them in plain language.
- 05Planned
Execution
Approved algorithms running around the clock, with monitoring.
- 06Planned
Community & marketplace
Research sharing, leaderboards and strategy discovery.
Backtesting
In buildBacktests that price in real trading.
| Capability | What it does |
|---|---|
| Historical data | Run strategies over chosen periods and assets |
| Transaction costs | Model commissions, fees and trading costs |
| Slippage | Gap between theoretical and executable price |
| Liquidity constraints | No unrealistic fills or oversized positions |
| Look-ahead controls | No future data leaking into past decisions |
| Survivorship-bias controls | Include delisted assets, not just survivors |
| Walk-forward testing | Train on one period, test on a later one |
| Out-of-sample testing | Measure on data the strategy never saw |
| Parameter sensitivity | Check results don’t hinge on one setting |
| Monte Carlo + robustness | Stress outcomes across markets and periods |
Historical data
- What it does
- Run strategies over chosen periods and assets
Transaction costs
- What it does
- Model commissions, fees and trading costs
Slippage
- What it does
- Gap between theoretical and executable price
Liquidity constraints
- What it does
- No unrealistic fills or oversized positions
Look-ahead controls
- What it does
- No future data leaking into past decisions
Survivorship-bias controls
- What it does
- Include delisted assets, not just survivors
Walk-forward testing
- What it does
- Train on one period, test on a later one
Out-of-sample testing
- What it does
- Measure on data the strategy never saw
Parameter sensitivity
- What it does
- Check results don’t hinge on one setting
Monte Carlo + robustness
- What it does
- Stress outcomes across markets and periods
Live trade intelligence
PlannedEvery trade gets a thesis, then a verdict.
Presented as analysis of observed data, not certainty about causation.
- 01
Detects a new trade from your connected exchange
- 02
Analyses size, exposure, volatility, momentum and regime
- 03
Suggests a thesis; you confirm or correct it
- 04
Compares the thesis with the outcome
- 05
Finds patterns across your history
Compared
Where Amstos fits.
| Capability | Crypto trackers | Trading journals | Amstos |
|---|---|---|---|
| Crypto exchange ingestion | Yes | Partial | Yes |
| Risk analytics (VaR, CVaR, Sharpe, Sortino) | No | Partial | Yes |
| Attribution & factor exposure | No | Partial | Yes |
| Regime & anomaly detection | No | Partial | Yes |
| Portfolio optimisation | No | No | Yes |
| Scenario & stress testing | No | No | Yes |
| Walk-forward + Monte Carlo backtesting | No | Partial | Yes |
| Live trade intelligence | No | No | Yes |
| AI behavioural coaching on every tier | No | Partial | Yes |
| Multi-portfolio analysis | No | Partial | Yes |
Crypto exchange ingestion
- Crypto trackers
- Yes
- Trading journals
- Partial
- Amstos
- Yes
Risk analytics (VaR, CVaR, Sharpe, Sortino)
- Crypto trackers
- No
- Trading journals
- Partial
- Amstos
- Yes
Attribution & factor exposure
- Crypto trackers
- No
- Trading journals
- Partial
- Amstos
- Yes
Regime & anomaly detection
- Crypto trackers
- No
- Trading journals
- Partial
- Amstos
- Yes
Portfolio optimisation
- Crypto trackers
- No
- Trading journals
- No
- Amstos
- Yes
Scenario & stress testing
- Crypto trackers
- No
- Trading journals
- No
- Amstos
- Yes
Walk-forward + Monte Carlo backtesting
- Crypto trackers
- No
- Trading journals
- Partial
- Amstos
- Yes
Live trade intelligence
- Crypto trackers
- No
- Trading journals
- No
- Amstos
- Yes
AI behavioural coaching on every tier
- Crypto trackers
- No
- Trading journals
- Partial
- Amstos
- Yes
Multi-portfolio analysis
- Crypto trackers
- No
- Trading journals
- Partial
- Amstos
- Yes
✓ included · ~ partial · — not included
FAQ
Common questions
How many engines are there?
18 at full build, rolled out in phases: 4, then 9, then 18.
Does the AI make things up?
It explains results the engines already calculated, so every statement traces back to a number.
Is my data safe?
Amstos only reads your trade history and never resells it.
How far back does backtesting go?
Up to 8 yrs of crypto history, depending on plan.