Last updated: February 2026 • Latest Version analysis
⚠️ Searching for NOVA s7 EA free download? Free versions of premium EAs are either outdated, decompiled fakes, or bundled with malware. NOVA s7 is an official MQL5 marketplace product — cracked .ex5 files do not exist for marketplace EAs. Sites offering “free downloads” are scams that deliver password-locked zip files leading to Telegram payment traps. We offer the lowest legitimate price available — with a 7-day money-back guarantee.
🚀 Ready to Buy NOVA s7 EA? Save ~50% vs MQL5
MT5: $239.95 (MQL5: $477)
✅ DeepSeek AI-powered Gold (XAUUSD) trading system — adaptive multi-layer intelligence
✅ +1,997% verified MyFXBook gain over 15 months — $800 → $16,781
✅ MQL5 Signal #2360172 live — +205% growth, 19.68 profit factor, 3.48% max drawdown
✅ 14.93 profit factor and 98.9% win rate across 2,906 backtest trades
✅ Perfect 5★ MQL5 rating — by the creator of META i7, META i9, and META i11
✅ Contextual AI that adapts to current market conditions in real time
30 Second Overview of NOVA s7 EA
NOVA s7 EA is a fully automated Gold (XAUUSD) trading system for MetaTrader 5 built around a DeepSeek AI framework. Developed by Meta Sophie Agapova (MQL5 username: metasphere), it trades XAUUSD by interpreting market behaviour contextually rather than reacting to fixed indicator signals. The developer describes it as an “Institutional Adaptive AI Trading Engine” that continuously evaluates market structure, momentum shifts, volatility pressure, and execution quality through a multi-layer intelligence system.
The philosophy: Unlike traditional EAs that rely on static indicator rules (such as “buy when RSI drops below 30”), NOVA s7 attempts to process market context through an AI framework — evaluating what the market is doing right now and adjusting its approach dynamically. The developer claims the system uses DeepSeek AI for contextual decision-making, though the specific implementation details of how this AI layer integrates with the MQL5 execution engine are not publicly documented.
The developer: Meta Sophie Agapova is a relatively new presence on the MQL5 marketplace, with her first product (META i7) published in November 2025. Despite the short track record, her portfolio has accumulated strong ratings: META i11 (5★, 17 reviews), META i9 (5★, 7 reviews), META i7 (5★, 4 reviews), and NOVA s7 (5★, 4 reviews). Reviewers across her products consistently praise her responsiveness and hands-on installation support. Multiple buyers report owning 4+ EAs from her — a notable signal of customer retention.
The numbers: The headline claim is +1,997% verified MyFXBook gain over 15 months from an $800 initial deposit, averaging 22% monthly returns. The developer also launched an MQL5 Signal (#2360172) in February 2026 showing +205% growth from $1,200 in the first 3 weeks (10 trades, 90% win rate, 19.68 profit factor). The backtest across 2,906 trades produces a 14.93 profit factor, 92.64 recovery factor, and 98.9% win rate — with a near-perfect 1.00 LR correlation indicating linear equity growth. MQL5 retail price is $477. The EA currently holds a perfect 5★ rating across 4 MQL5 reviews.
Our call: NOVA s7 has impressive numbers and a developer who clearly inspires loyalty among her customers. The DeepSeek AI claims are intriguing, and the live MyFXBook result is attention-grabbing. The developer has also launched an MQL5 Signal, though it’s only 3 weeks old with 10 trades — too early to draw conclusions. This is a very new product from a very new developer — the entire portfolio launched within the last few months. The 98.9% win rate combined with a 1:5 reward-to-risk ratio (average win $30K vs average loss $153K) means when losses come, they’re significant. The backtest covers only 13 months. At $239.95 through CheaperForex (vs $477 on MQL5), you’re saving ~50% and can test with lower financial commitment while the EA builds a longer track record.
Rating: 3/5 ⭐⭐⭐ — Promising AI-Powered Gold System With Impressive Claims But Very Limited Track Record

The Developer: Meta Sophie Agapova
Meta Sophie Agapova develops AI-driven trading systems under the MQL5 username metasphere. Her portfolio focuses on neural network and AI-powered Expert Advisors, with a clear progression from dual neural networks (META i7) to quantum adaptive systems (META i9) to hybrid cognitive architectures (META i11), and now DeepSeek AI integration (NOVA s7).
Developer Portfolio:
• META i11 — Hybrid Cognitive Trading System (5★, 17 reviews) — her most-reviewed product
• NOVA s7 — DeepSeek AI Gold Trading Engine (5★, 4 reviews) — this product
• META i9 — Quantum Adaptive Trading Engine (5★, 7 reviews)
• META i7 — Dual Neural Network EA (5★, 4 reviews)
What stands out: Every product in her portfolio holds a perfect 5.0★ rating. Multiple reviewers explicitly state they own 4+ EAs from this developer, with one describing them as “all winners.” MQL5 reviewers consistently highlight her responsiveness and willingness to assist with installation and configuration. One reviewer noted profitable results across both demo and real accounts after just two weeks.
The context you need: The entire portfolio launched within the last few months (META i7 was published November 2025). While the perfect ratings and customer loyalty are genuinely positive signals, the overall track record is still very young compared to established developers with years of marketplace history. The 32 combined reviews across all products are encouraging but don’t yet represent the depth of community validation you’d see with long-established EAs.
How NOVA s7 Trades
NOVA s7 is described as an Institutional Adaptive AI Trading Engine that processes Gold (XAUUSD) price action through multiple analytical layers simultaneously:
Multi-Layer Intelligence System:
1. Market Structure Analysis — Evaluates the current structural environment rather than relying on fixed chart patterns
2. Momentum Shift Detection — Identifies changes in directional pressure before they become visible through traditional indicators
3. Volatility Pressure Assessment — Adapts position management based on current volatility conditions
4. Execution Quality Monitoring — Evaluates trade execution quality in real time and adjusts accordingly
The core differentiator, according to the developer, is that NOVA s7 makes contextual decisions — adjusting its trading approach based on what the market is doing right now rather than applying the same rules regardless of market conditions. The DeepSeek AI framework is claimed to process these layers simultaneously and generate trading decisions that account for the current state of all four analysis dimensions.
What this likely means in practice: The EA appears to use an AI/machine learning model to classify market conditions and select appropriate trading parameters dynamically. Rather than a fixed “buy when X and Y happen” approach, the system likely assigns probability weights to different market states and adjusts entry timing, position sizing, and exit management based on the assessed environment. The 98.9% win rate suggests the system is highly selective about entries and uses wide stop losses relative to take-profit targets.
Understanding the risk profile: The backtest data reveals an average win of $30,178 versus an average loss of -$153,573 — roughly a 1:5 reward-to-risk ratio. This means when the EA wins, it wins modestly; when it loses, it loses approximately 5x the typical win. The 98.9% win rate compensates for this asymmetry, but traders should understand that a single loss can erase multiple winning trades. This is characteristic of high-win-rate strategies that use wider stop losses — it works well until market conditions trigger a cluster of losses.
Live Performance, Backtests & Reviews
Verified MQL5 Live Signal
The developer maintains a verified live signal on MQL5 Signals — Signal #2360172. This signal launched very recently (February 19, 2026) and has limited history, but it does provide broker-level verification through MQL5’s infrastructure:

| Metric | Value |
|---|---|
| Signal | #2360172 — NOVA s7 |
| Growth | +205% |
| Tracking Period | 3 weeks (started Feb 19, 2026) |
| Initial Deposit | $1,200 |
| Balance / Equity | $3,665 |
| Profit | $2,465 |
| Total Trades | 10 |
| Win Rate | 90% (9 wins, 1 loss) |
| Profit Factor | 19.68 |
| Recovery Factor | 18.68 |
| Best / Worst Trade | +$728 / -$132 |
| Avg Holding Time | 57 minutes |
| Max Balance Drawdown | 3.48% ($132) |
| Max Equity Drawdown | 3.41% ($129) |
| Broker / Leverage | Weltrade / 1:500 |
| Algo Trading | 100% |
⚠️ Important context on the signal: This MQL5 Signal is only 3 weeks old with 10 trades. MQL5 itself flags multiple warnings: “too much growth indicates high risk,” “newly opened account — results may be random,” and “number of deals too small to evaluate trading quality.” The 205% growth is eye-catching but statistically meaningless over such a small sample. The signal needs months of consistent trading before it can be considered a reliable performance indicator. The developer also references a longer-running MyFXBook account showing +1,997% gain over 15 months from an $800 deposit, which provides additional (though separately verified) context.
Backtest Results (Jan 2025 – Feb 2026)
The developer provides a comprehensive backtest using dynamic lot sizing:


| Metric | Value |
|---|---|
| Backtest Period | January 2025 – February 2026 (13 months) |
| Initial Deposit | $5,000 |
| Total Net Profit | $80,921,942.55 |
| Profit Factor | 14.93 |
| Recovery Factor | 92.64 |
| Sharpe Ratio | 8.79 |
| LR Correlation | 1.00 (perfect linear growth) |
| Total Trades | 2,906 |
| Win Rate | 98.90% (Long: 98.84% | Short: 99.01%) |
| Average Win / Average Loss | $30,178 / -$153,573 (1:5 ratio) |
| Max Consecutive Wins | 400 trades ($14.2M) |
| Max Consecutive Losses | 2 trades (-$420,700) |
| Balance Drawdown Maximum | 2.28% |
| Equity Drawdown Relative | 21.38% |
| History Quality | 99% |
⚠️ Backtest reality check: The $80.9M profit from a $5,000 deposit looks extraordinary, but this uses dynamic lot sizing — meaning position sizes compound aggressively as the account grows. This dramatically inflates the headline profit number. A fixed-lot backtest would show far more modest (but realistic) numbers. Additionally, the 13-month backtest period is very short and covers only a recent Gold bull run. A robust backtest should span multiple years including both bull and bear Gold conditions. Treat these numbers as indicative of strategy potential, not expected real-world returns.
Review Analysis (4 Verified MQL5 Reviews at 5★)
NOVA s7 currently has 4 reviews on MQL5, all 5-star. While the review count is still small, we can cross-reference with the developer’s other products — particularly META i7 and META i9 — which share the same developer and similar architecture. Combined across all products, the developer has 32 reviews. Common themes include:
What reviewers consistently say (across developer’s portfolio):
✅ Profitable results on both demo and real accounts within the first weeks
✅ Strong risk-reward management on individual trades
✅ Developer is professional, responsive, and provides hands-on setup assistance
✅ Multiple buyers own 4+ EAs from this developer — described as “all winners”
✅ Setup is straightforward with developer guidance
What we don’t yet have: With only 4 NOVA s7-specific reviews, there’s insufficient data to identify patterns around drawdown experiences, long-term performance, or how the EA handles adverse Gold conditions. The 32 combined reviews across the portfolio are positive but still represent a small sample compared to established EAs with hundreds of reviews.
Settings & Requirements
| Requirement | Details |
|---|---|
| Platform | MetaTrader 5 only (no MT4 version) |
| Instrument | XAUUSD (Gold) |
| Account Type | ECN/Raw Spread recommended for Gold trading |
| Recommended Deposit | $500+ (based on live account starting at $800) |
| VPS | Recommended for 24/5 uninterrupted operation |
| Broker | Low-spread Gold broker (IC Markets, Pepperstone, etc.) |
Gold-specific considerations: XAUUSD is one of the most volatile instruments available on retail platforms. Gold spreads vary significantly between brokers, and widened spreads during news events or low-liquidity sessions can significantly impact a high-frequency Gold EA. Choose a broker with consistently tight Gold spreads and fast execution. The live MyFXBook result appears to be on a low-spread broker environment.
How to Download NOVA s7 EA
There are two ways to get NOVA s7 EA for MetaTrader 5:
Option 1: Buy from MQL5 Marketplace (Full Price: $477)
Purchase directly from the official MQL5 marketplace listing. You pay the full retail price and receive the EA via MQL5’s standard activation system.
Option 2: Buy from CheaperForex (Discounted: $239.95 — Save ~50%)
CheaperForex purchases licenses in bulk and passes the savings to you. You get the exact same official MQL5 product — activated directly from the marketplace — at a significantly lower price. Free installation, free lifetime updates, and a 7-day money-back guarantee before activation.
What You Receive:
✅ Latest Version of NOVA s7 EA — installed directly from MQL5
✅ Free lifetime updates — downloaded automatically within MT5
✅ Presets included (if available from the developer)
✅ Free installation support — CheaperForex will install it for you if needed
💰 Save ~50% on NOVA s7 EA MT5
Get the exact same official MQL5 product at a fraction of the price. Same EA, same updates, same lifetime license.
Frequently Asked Questions
What is NOVA s7 EA and what does it trade?
NOVA s7 EA is a fully automated Expert Advisor for MetaTrader 5 that trades Gold (XAUUSD). It uses a DeepSeek AI framework to interpret market behaviour contextually through multiple analytical layers — market structure analysis, momentum shift detection, volatility pressure assessment, and execution quality monitoring. The result is a system that adapts its trading approach based on current market conditions rather than applying fixed indicator rules.
What is the DeepSeek AI framework?
DeepSeek is a Chinese AI company known for its large language models and reasoning capabilities. The developer claims NOVA s7 is built around a DeepSeek AI framework that enables contextual market interpretation. The specific technical details of how DeepSeek AI integrates with the MQL5 trading engine are not publicly documented. As with any AI claims in EA marketing, the proof is in the verified live performance rather than the architectural description.
Is the 1,997% live gain verified?
The 1,997% gain is verified through MyFXBook, which connects directly to the broker’s trade server to confirm results. This provides a reasonable level of verification. Additionally, the developer launched an MQL5 Signal (#2360172) in February 2026, which provides broker-level verification through MQL5’s infrastructure — the gold standard for EA performance monitoring. However, the MQL5 Signal is only 3 weeks old with 10 trades, so it’s too early to treat it as meaningful performance data. MQL5 itself flags the signal with warnings about the small sample size and high-risk growth rate.
Why is the win rate 98.9% but the average loss is 5x the average win?
This is a deliberate strategy design choice common in high-win-rate systems. The EA uses relatively wide stop losses compared to its take-profit targets, which means most trades close profitably (98.9%) but the occasional loss is significantly larger than the typical win ($153K loss vs $30K win). The math works as long as the win rate stays high — but if market conditions change and the win rate drops even a few percentage points, the impact of those large losses compounds quickly. This is an important risk factor to understand before trading live.
How much money do I need to start?
The developer’s MyFXBook signal started with $800. For Gold trading with an AI system, a minimum of $500 is reasonable with very conservative lot sizing. However, $1,000+ provides better margin headroom during drawdown periods. Use the demo version first to understand how the EA behaves with your broker’s Gold spreads and execution quality before committing real capital.
Is NOVA s7 available for MT4?
No. NOVA s7 is MT5-only. There is no MetaTrader 4 version available. The AI framework and multi-layer processing likely require MT5’s enhanced capabilities compared to the older MT4 platform.
Why is the backtest period only 13 months?
The backtest covers January 2025 to February 2026. This is a relatively short testing period. Most experienced traders prefer to see backtests spanning at least 3–5 years to include various market conditions — including both Gold bull runs and significant corrections. A 13-month backtest during what has been a strong Gold period may not represent how the EA performs during sustained Gold sell-offs or sideways consolidation. Demo testing across several months of live market conditions is strongly recommended before deploying real capital.
Can I use NOVA s7 on a prop firm account?
The EA’s backtest shows a 2.28% maximum balance drawdown and 21.38% equity drawdown. The balance drawdown figure may technically fall within prop firm daily drawdown limits, but the equity drawdown figure is concerning for strict prop firm rules. Always test on a demo account that mirrors your prop firm’s conditions before attempting funded trading. Many prop firms have specific rules about Gold trading, position sizing, and drawdown that you’ll need to verify.
Is a free download of NOVA s7 EA available?
No. NOVA s7 is an official MQL5 marketplace product that requires activation through the marketplace. Cracked .ex5 files do not exist for marketplace products. Sites advertising free downloads are scams — they typically deliver password-locked archives that lead to Telegram payment requests for junk files. CheaperForex offers the genuine product at ~50% below MQL5 retail price with official marketplace activation and a 7-day money-back guarantee.
Our Assessment
Strengths:
✅ +1,997% verified MyFXBook gain from $800 deposit over 15 months
✅ MQL5 Signal #2360172 launched — +205% in first 3 weeks, 19.68 profit factor, 3.48% max drawdown
✅ Perfect 5★ rating across all developer products (32 combined reviews)
✅ DeepSeek AI framework offers adaptive rather than static trading logic
✅ 14.93 profit factor and 92.64 recovery factor on backtest
✅ Developer is responsive, provides hands-on installation support, retains loyal customers
✅ 98.9% win rate provides psychologically comfortable trading experience during winning streaks
Concerns:
⚠️ Very new product from a very new developer — entire portfolio launched within last few months
⚠️ Only 4 MQL5 reviews for NOVA s7 — insufficient for pattern identification
⚠️ MQL5 Signal only 3 weeks old — 10 trades total, MQL5 flags “results may be random” and “too few deals to evaluate”
⚠️ 13-month backtest only — doesn’t cover sustained Gold bear conditions
⚠️ 1:5 reward-to-risk ratio — average loss of $153K vs $30K average win means losses are devastating
⚠️ Dynamic lot sizing inflates backtest headline numbers — $80.9M from $5K is not realistic expectation
⚠️ “DeepSeek AI” claims are unverifiable — no way to confirm what AI architecture is actually running inside the EA
⚠️ 21.38% equity drawdown on backtest — despite the 98.9% win rate, drawdowns are real
Bottom line: NOVA s7 is an intriguing product with genuinely impressive live results and a developer who clearly cares about her customers. The 1,997% MyFXBook gain is real and verified at the broker level. The recent launch of MQL5 Signal #2360172 is a positive step toward full transparency, though at only 3 weeks old with 10 trades it’s too early to draw conclusions. The problem is time — this product and developer are both very new, and the spectacular numbers come from a short track record during favourable Gold conditions. The 98.9% win rate with 5x loss ratio is a strategy design choice that works brilliantly until it doesn’t. We’d rate this higher with 6+ months of MQL5 Signal data and multi-year backtests. For now, it deserves a cautious 3/5 — promising but unproven over the long term. At $239.95 through CheaperForex, the reduced price point makes testing substantially less risky than paying $477 on MQL5.
Recommendation: Start on demo for a minimum of 4–8 weeks. Monitor how the EA handles Gold volatility during news events and any adverse trending conditions. Pay close attention to the size and frequency of losses relative to wins. If the live performance aligns with the backtest metrics after extended demo testing, consider a small live account with conservative lot sizing. Do not expect 22% monthly returns on a live account — that figure reflects aggressive compounding from a small starting balance.
🚀 Get NOVA s7 EA MT5 — Save ~50% vs MQL5
$239.95 (MQL5: $477)
✅ DeepSeek AI Gold trading system — +1,997% verified MyFXBook gain
✅ Official MQL5 activation — not cracked or modified
✅ Free installation, free lifetime updates, 7-day money-back guarantee
Disclaimer: Forex and Gold trading involves substantial risk. Past performance — including the +1,997% MyFXBook result — does not guarantee future returns. The backtest figures use dynamic lot sizing which dramatically inflates headline numbers. A 98.9% win rate does not mean the EA cannot lose money — when losses occur, they are approximately 5x larger than typical wins. The 21.38% equity drawdown on backtest demonstrates that significant temporary losses are possible. Never trade with money you cannot afford to lose. CheaperForex is a reseller of official MQL5 marketplace products — we are not the developer and do not guarantee trading performance. Always demo test extensively before live deployment.