
Strategy workspace
Build, evaluate and challenge systematic trading ideas against your own market data.

Recent research
| Run | Dataset | Selected family | OOS return | Status |
|---|---|---|---|---|
| Loading research history… | ||||
Research queue
New strategy search
Choose source data and validation settings. The engine searches defined rule families and ranks candidates using in-sample data only.
Advanced settingsValidation · costs · Monte Carlo · parameter ranges
Sign in on the hosted website to see your plan’s genetic controls.
Market data
Inspect supplied MT5 Demo-Server OHLC files or import OHLC bars / bid-ask ticks.
Dataset inventory
| Dataset | Symbol | Timeframe | Type | Size | Preview |
|---|
Live Run
Monitor actual search status and completed stages for the current run.
> awaiting strategy search
Run health
Current candidate verdicts
Results
Review IS selection, unseen OOS performance, walk-forward folds and Monte Carlo distributions.
Saved runs
| Created (UTC) | Dataset | Status | Run | Open |
|---|
Portfolio
Archive and compare strategy candidates. Cross-asset portfolio sizing and combined capital curves are not enabled in this build.
Archived strategies
Research runs
Indicators & rule blocks
Only the calculation families wired into this search are selectable. The complete FSB 102-file traceability registry is not yet researched or implemented.
| Family | Parameters | Current use | Data needs | State |
|---|---|---|---|---|
| EMA crossover | fast / slow bars | Trend signal | Close | IN SEARCH |
| EMA + ADX trend following | fast / slow EMA / ADX threshold | Trend and strength filter | OHLC | IN SEARCH |
| RSI mean reversion | period / lower / upper | Mean-reversion signal | Close | IN SEARCH |
| Bollinger breakout | period / standard deviations | Breakout signal | Close | IN SEARCH |
| Donchian breakout | lookback bars | Range breakout | High / low / close | IN SEARCH |
| Supertrend (ATR trend) | ATR period / multiplier | Trend reversal state | OHLC | IN SEARCH |
| Keltner (ATR channel breakout) | EMA / ATR period / multiplier | Volatility channel breakout | OHLC | IN SEARCH |
| Chandelier (ATR trend reversal) | lookback / ATR / multiplier | Trailing trend reversal | OHLC | IN SEARCH |
| ATR impulse breakout | ATR period / threshold | Volatility-scaled momentum | OHLC | IN SEARCH |
| ATR pullback continuation | fast / slow EMA / ATR depth | Trend pullback continuation | OHLC | IN SEARCH |
| ATR expansion range breakout | range / ATR / expansion | Volatility expansion breakout | OHLC | IN SEARCH |
Robustness
Monte Carlo and WFO results are attached to each completed run. Choose a run to inspect its report.
Available runs
Correlation
Compare Pearson correlation on aligned OOS daily returns from two completed runs. Minimum sample: 30 shared dates.
MT5 Link
Pair a terminal account for identity and heartbeat only. Historical data import is a separate, explicit action.
Generate a code, download and attach the read-only EA in MT5, then enter the code in its inputs. Add to the terminal's WebRequest allow-list.
Import selected OHLC history from MT5
Only the selected symbol/time range is transferred after you request it. Import uses CopyRates; tick import is not available yet.
MQL5 Export
Generate an original EA only for supported rule families. Risk settings are explicit and live trading is disabled by default.
Affiliates
Share your SYNBAI Nexus link. Your audience receives 10% off every renewal; you earn 20% of each final payment for as long as their subscription stays active.
Fair value, shared long-term
Download your branded banners
| Date | Sale | Your 20% | Status |
|---|---|---|---|
| Sign in on fxsynbai.eu to view tracked commissions. | |||
Account & plan
Verified identity starts the trial once. Entitlements are stored and checked server-side.
MONEY-BACK
30-day money-back guarantee
Eligible first paid subscription purchase on an account; statutory consumer rights remain in force.
Read the full Refund Policy| Feature | Trial | Starter | Pro |
|---|---|---|---|
| Monte Carlo | ✓ 1/day · 15 total | ✓ 30/month | ✓ Unlimited |
| Pearson correlation | ✓ 1/day · 15 total | ✓ 30/month | ✓ Unlimited |
| MQL5 source export | ✓ 1 total | ✓ 1/day | ✓ Unlimited |
| Archived strategies | ✓ Up to 20 | ✓ Up to 50 | ✓ Unlimited |
| Island-based genetic scan | ✓ 3 islands | ✓ 3 islands | ✓ 1–10 islands |
| Population per island | ✓ 75, fixed | ✓ 75, fixed | ✓ 75–150 · step 5 |
| Crossover rate | ✓ 70%, fixed | ✓ 70%, fixed | ✓ Editable |
| Edit genetic controls | ✕ | ✕ | ✓ |
| Referral discount | ✓ Extra 10% when referred | ✓ Extra 10% when referred | ✓ Extra 10% when referred |
$0 / 15 days
Start with one Monte Carlo test and one Pearson correlation each day. Each tool has 15 total trial uses; unused daily uses expire.
- One strategy research run with Monte Carlo per day
- One Pearson correlation per day
- One MQL5 export total during Trial
- Trial ends after 15 days
$20 / month
For focused research with a monthly testing allowance.
- 30 Monte Carlo tests each month
- 30 Pearson correlations each month
- 1 MQL5 source export per day
$40 / month
For users who want every supported analysis available.
- Unlimited Monte Carlo tests
- Unlimited Pearson correlations
- Unlimited strategy exports and archive
$17 one-time upgrade
Upgrade an active Starter subscription for $17 now. First-month spend is $37 total ($20 Starter + $17 upgrade); following months renew at $40.
Settings
Configure SYNBAI Nexus, your AI workspace, and local application behavior.
Update feed
A research workspace for historical strategy analysis. It does not place trades. The AI model manager is being built; assistants remain unavailable until a supported inference runtime is connected.
Last updated: 14 November 2026
This page summarizes how SYNBAI Nexus handles information in its current local and hosted modes. The public hosted service is not yet deployed; its final policy and contact details must be published before launch.
Local desktop use
Imported market files, research runs, results, exports, and preferences are stored by the local application on this device. General preferences and AI agent selections are saved in this browser profile. Nexus AI project names, uploaded files, and supported text extracted from them are stored in local IndexedDB on this device; files are not sent to a provider in this build. Opening Settings → Models requests the public GGUF catalog from Hugging Face, and model searches are sent to that service. Do not enter API secrets here; this settings screen does not store them.
Hosted account and research
When hosted features are available, account details such as email address, verification status, subscription entitlement, and usage records are processed by the SYNBAI service. Uploaded datasets, research requests, results, and archived strategies are processed to provide the service. Retention periods will be disclosed in the complete policy before hosted launch. Billing is handled by the selected payment provider; SYNBAI does not receive full payment-card details.
MT5 connection
The MT5 link is read-only. It does not read trading passwords, balances, equity, positions, orders, or trade history. The service stores a keyed hash of the account login and connection metadata needed to maintain the link. Market history is transferred only when the user requests an import.
AI providers
No AI provider receives your prompts or research data through this build: AI inference is not connected yet. When AI is enabled, the selected provider, the information sent, and the applicable retention will be shown before use. Cloud-provider credentials must be handled through a protected integration and must never be saved in this page.
Your choices
You can choose whether to use hosted features, remove imported datasets from the Data page, and select or clear local AI preferences. Before public launch, the complete policy will identify the data controller, contact channel, retention periods, processors, and applicable privacy-rights process.
Choose where local inference will run. Model profiles are saved locally; no provider API key is stored here.
Live public catalog from Hugging Face Hub. Results are paged and include all public repositories tagged GGUF by the Hub. Review each model card and license before use.
Create focused assistants for your workflow. Profiles and permission choices stay in this browser profile. An installed model is selected separately under Models.
Side-effect actions are off by default. These are preferences only until an AI runtime is connected.
User Manual
A beginner-friendly route from market data to a strategy you can inspect. The app researches historical data; it never places trades.
Start with XAUUSD
Open Data, start with the supplied XAUUSD M15 demo dataset, or import your broker’s MT5 CSV. M15 offers more observations and more noise than higher timeframes.
Use a preset first
Go to New Strategy and choose Automatic XAUUSD scan to compare supported rule families, or select one of the focused EMA/ATR templates. You can leave Advanced settings closed.
What the checks mean
The search ranks candidates using the in-sample segment, then measures a held-out out-of-sample segment. Walk-forward analysis repeats this over time windows. Monte Carlo resamples completed OOS trades to show how sensitive the sample may be to trade ordering.
Correlation and MQL5
Pearson correlation compares aligned daily returns from two completed runs; it is not a strategy score. Archive a candidate only after reviewing its costs, drawdown, trade count, OOS and WFO results. MQL5 exports are source code and must be compiled and demo-tested in MetaEditor/MT5.
Download broker history from MT5
- In MT5, open Market Watch. Right-click inside it and choose Symbols.
- Find your broker’s gold symbol (it may be named XAUUSD or have a suffix), open Bars, choose the timeframe and date range, then click Request.
- When the history loads, click Export and save the CSV somewhere you can find it.
- In SYNBAI Nexus, open Data → Import CSV, choose that file, review the preview, then run research on the imported dataset.
Small glossary
- Strategy
- A set of entry and exit rules that turns market bars into hypothetical positions.
- IS / OOS
- In-sample data is used to select a candidate. Out-of-sample data is held back for a separate historical check.
- WFO
- Walk-forward analysis repeats selection and evaluation over sequential time windows; it does not prove future performance.
- Monte Carlo
- A statistical resampling of completed historical trades. It is not a tick replay or a forecast.
- Pearson correlation
- A number from −1 to +1 describing the linear relationship between two aligned return series. Correlation does not establish diversification by itself.


