What is StratBase.ai?
StratBase.ai is a web-only, subscription SaaS backtesting tool founded in 2024 and operated by a small team described as roughly six people, listed by first names only, based in Kyiv, Ukraine. The vendor names Ruslan Vlasiuk as founder and CEO, though we could not independently verify that, and the company LinkedIn it links to belongs to an unrelated CS:GO entity. There is no legal entity name, no registration number, and no physical address published anywhere on the site. The governing-law clause in the terms does not even name a jurisdiction, referring only to “the laws of the jurisdiction where StratBase.ai operates.”
Functionally, it is an automated backtester. You build a strategy, either through the AI natural-language builder or by assembling rules from a library the vendor states includes 245+ indicators, then run it bar by bar over historical data for forex, crypto, or US stocks. It does not do manual visual replay, it has no MT4, MT5, cTrader, or NinjaTrader integration, no desktop app, and no mobile app. It is a browser product and nothing else.
Our verdict
StratBase.ai is a capable-looking but unproven backtester carrying an unusually heavy ownership risk. The engineering on paper is real and the free tier is genuinely useful, but we cannot recommend that anyone build a paid workflow on a product that is publicly advertised for sale, has no named legal entity, has zero independent reviews, and whose terms say nothing about what happens to your subscription or data if it is sold or shut down. The capability is promising. The continuity is not. We score it 5.4 / 10.
We assess StratBase.ai from its public pages, its own blog, pricing, terms, and refund policy, and a search for independent regulator, marketplace, and user-sentiment signals. We have not purchased a subscription or run a backtest.
Key features & specs
The headline is an AI strategy builder powered by Anthropic's Claude, which converts instructions like “buy when RSI crosses above 30” into executable rules. Around that sit grid and DCA templates, multi-timeframe support, grid-search and walk-forward optimisation, AI-written analysis reports, a proprietary “StratBase Score,” and an output suite covering equity curve, drawdown, underwater chart, trade log, and a set the vendor calls “20+ metrics” including Sharpe, Sortino, Calmar, max drawdown, and win rate. It runs only in the browser, with no MT4/MT5/cTrader/NinjaTrader, no desktop or mobile, and no manual replay. Data comes from named vendors (Dukascopy for forex, Binance and Bybit for crypto, Alpaca for US stocks), but the forex data is Dukascopy tick aggregated to one-minute candles despite “tick-level” marketing, and execution modelling for slippage, spread, and swaps is contradictory or undocumented. Note what is absent or unconfirmed: Monte Carlo is not mentioned anywhere and should be treated as not shipped, and the export format for reports is unconfirmed.
| Attribute | Value | Source |
|---|---|---|
| Company & track record | Operated by “StratBase.ai” with no legal entity name, registration number, or physical address disclosed anywhere. The vendor names Ruslan Vlasiuk as founder/CEO and lists a team of roughly six people by first name only, based in Kyiv, Ukraine, with a stated founding year of 2024. We could not independently verify the founder (the linked company LinkedIn belongs to an unrelated CS:GO entity), and there are no Wayback snapshots. Most strikingly, a sitewide banner advertises the entire business for sale: “StratBase.ai is for sale — $20,000,” with a “Learn more” link that returns a 404 and no public marketplace (Acquire/Flippa) listing we could find. | StratBase.ai ↗ |
| Software status | A backtesting web app, not a broker, adviser, or funded-trader program. It holds no financial licence and is not financially regulated, which is normal and expected for backtesting software. Its terms state plainly that “StratBase.ai is not a financial advisor, broker, or investment service” and that “AI does NOT provide trading recommendations or financial advice.” No regulator warning naming the product was found. The fair criticism here is corporate opacity, not regulation: the governing-law clause does not even name a jurisdiction (“the laws of the jurisdiction where StratBase.ai operates”). | StratBase.ai terms ↗ |
| Pricing & plans | Subscription SaaS, monthly, with no lifetime or one-time option. The free tier is permanent (5 backtests/day, 1-minute minimum timeframe, up to 1 year of data). Paid plans are Pro at $29/mo (unlimited backtests, 3 years of data, single-parameter optimisation), Premium at $49/mo (1-second resolution, 5 years of data, full multi-parameter optimisation), and Private at $99/mo (confidential strategy visibility, priority support). A “yearly, save ~28%” toggle exists, but the exact annual prices are not published. Payments run through Paddle, with NowPayments for crypto. | StratBase.ai pricing ↗ |
| Free tier & refund | The free tier is permanent and meaningful, and the vendor describes it as “an unlimited trial — use it as long as you want,” so there is no separate timed trial to auto-convert. Paid plans carry a genuine “14-day unconditional money-back guarantee on all purchases,” processed via Paddle as Merchant of Record with a 24-hour response target and 5-10 business days to land. Auto-renewal is disclosed in plain language and cancellation is self-serve “anytime from your dashboard — no questions asked.” | StratBase.ai terms ↗ |
| Platform & compatibility | Browser-based cloud web app only. There is no MT4, MT5, cTrader, or NinjaTrader plugin, no desktop build, no native mobile app, and no manual visual replay mode. The tech stack is a Rust engine with Python/PyO3, a Next.js frontend, PostgreSQL with TimescaleDB, and Anthropic Claude for AI features. Data is pre-loaded and bundled; “tools” pages reference optional CCXT/Polygon import, but a native Polygon integration is claimed only on a tools page rather than in documentation, so we treat import as unverified. | StratBase.ai ↗ |
| Data quality & sources | The vendor names real, reputable sources: Dukascopy for forex, Binance and Bybit for crypto, and Alpaca for US stocks, with coverage of 27 forex pairs, 1700+ crypto pairs (spot and futures), and roughly 130 to 230 US stocks, and depth claimed at “5+ years.” The headline oversells it, though: the site markets “tick-level source data,” but the vendor's own blog admits the forex data is Dukascopy tick data “aggregated to 1-minute candles.” For forex you are backtesting on M1 bars, not ticks. The finest resolution, 1-second candles, applies to crypto on Premium and above. Of eight exchange logos shown, only Binance, Bybit, Dukascopy, and Alpaca are confirmed functional, and there is no independent data audit. | StratBase.ai blog ↗ |
| Execution modelling | The vendor states a “next-bar execution” rule with “indicators on completed bars only” to limit look-ahead bias, which we cannot independently audit. Commission appears to use pre-loaded maker/taker fee schedules for Binance and Bybit. Beyond that the modelling is murky: slippage is described two contradictory ways across the vendor's own content (user-entered constants in one blog, “dynamic slippage by volatility/time of day” of 0.02-0.08% in another), and neither appears in documentation. Forex spread handling (fixed vs variable) and swap/funding application are undocumented, and partial fills, requotes, and weekend gaps are not mentioned at all. | StratBase.ai blog ↗ |
| Mode, features & honesty | Automated bar-by-bar backtesting only, with no manual replay. The standout shipped feature is an AI natural-language strategy builder powered by Anthropic Claude that turns instructions like “buy when RSI crosses above 30” into executable rules, alongside 245+ indicators across six categories, grid/DCA templates, multi-timeframe support, grid-search and walk-forward optimisation, AI-written analysis reports, a proprietary “StratBase Score,” and outputs covering equity curve, drawdown, underwater chart, trade log, and a set the vendor calls “20+ metrics.” Monte Carlo is not mentioned anywhere and should be treated as not shipped; export format is unconfirmed. On honesty the core disclaimers are good (past performance does not guarantee future results; results are simulations that may not reflect real-world conditions), and the vendor's own blog carries an overfitting warning. But headline accuracy claims are unbacked and self-contradictory, and supporting blog content runs under unverifiable author personas. | StratBase.ai ↗ |
Pricing & value
StratBase.ai is subscription-only and monthly. There is no lifetime or one-time option, which, given the for-sale banner, is the safer model for a buyer anyway. The free tier is permanent and meaningful: five backtests per day, a one-minute minimum timeframe, and up to one year of data. Paid tiers are Pro at $29/mo (unlimited backtests, three years of data, single-parameter optimisation), Premium at $49/mo (one-second resolution, five years of data, full multi-parameter optimisation), and Private at $99/mo (confidential strategy visibility and priority support). A “yearly, save ~28%” toggle exists, but the exact annual prices are not published, which is a transparency gap.
On value, the honest comparison is against free tools. The built-in MT4 and MT5 strategy testers cost nothing and remain the default for testing expert advisors on those platforms. TradingView's bar replay is free to a point and far better for discretionary visual practice, something StratBase does not do at all. So StratBase is competing on its AI builder, its Rust speed, and its pre-loaded multi-asset data, not on price. That can be worth $29 to $49 a month to the right user, but the for-sale risk means we would only spend it month to month, never annually.
Subscription only, monthly (yearly save ~28%, exact annual price not published); permanent free tier; 14-day unconditional refund via Paddle; no lifetime| Plan | Price | Billing | What's included | Best for |
|---|---|---|---|---|
| Free | $0 | Permanent | 5 backtests/day, 1-min min, up to 1yr data, 245+ indicators, all markets | Trying the engine at no cost |
| Pro | $29/mo | Monthly (yearly ~28% off) | Unlimited backtests, up to 3yr data, single-param optimisation | Active solo backtesters |
| Premium | $49/mo | Monthly (yearly ~28% off) | +1-second resolution, up to 5yr data, full multi-param optimisation | Crypto users wanting finest resolution |
| Private | $99/mo | Monthly (yearly ~28% off) | +confidential strategy visibility, priority support | Users guarding proprietary strategies |
Transparency & trust
A backtest is only as honest as the data and the fill model behind it, so this is one of the decisive parts of the review. The vendor names real, reputable sources: Dukascopy for forex, Binance and Bybit for crypto, and Alpaca for US stocks. Coverage is stated as 27 forex pairs, 1700+ crypto pairs, and roughly 130 to 230 US stocks, with depth claimed at “5+ years.” That sourcing is a genuine positive. The problem is the headline. The site markets “tick-level source data,” but the vendor's own blog admits the forex data is Dukascopy tick data aggregated to one-minute candles. That is an important distinction. For forex, you are backtesting on M1 bars, not on ticks, regardless of the marketing language. The finest resolution available, one-second candles, applies to crypto on Premium and above, not to forex. Eight exchange logos appear on the site, but only four, Binance, Bybit, Dukascopy, and Alpaca, are confirmed functional, and there is no independent audit of any of it.
On execution modelling, the vendor states a “next-bar execution” rule and says indicators are evaluated only on completed bars to avoid look-ahead bias. We cannot confirm that independently. Commission appears to be handled via pre-loaded maker and taker fee schedules for Binance and Bybit. After that, the modelling gets murky. Slippage is described two contradictory ways across the vendor's own content: one blog treats slippage and spread as user-entered constants, another claims “dynamic slippage by volatility and time of day” of 0.02 to 0.08 percent. Neither appears in documentation. Forex spread handling, fixed versus variable, and swap or funding application are undocumented. Partial fills, requotes, and weekend gaps are not mentioned at all. For a forex backtester specifically, that is a lot of unanswered execution detail.
“14-day unconditional money-back guarantee on all purchases.”
StratBase.ai Terms of Service, Refund Policy (Terms Sec 7), accessed June 2026 TOS ↗
Now the advertised results, which a reader should treat as marketing, not evidence. The vendor claims backtests deviate from live by “±0.3%,” though the same article also says “±0.4%,” an internal inconsistency that undercuts the figure. It cites a “64.2% backtest versus 63.8% live” BTCUSDT example, but that result is self-authored and self-graded, with no reproducible method or third-party audit. Claims that “competitors are off 5 to 15%” and that the engine runs years of data “in seconds” are similarly unverified. More troubling, the supporting blog content is published under author personas, Sarah Chen, James Mitchell, and David Ross, that we could not verify as real people and that read as house or AI bylines. The same content carries a fabricated Ed Seykota quote and an unverifiable “$180k saved” anecdote, both of which we reject and do not repeat as fact. None of this means a single advertised number is false. It means none of it can be trusted as evidence, and a good backtest never guarantees live results anyway, because curve-fitting and overfitting are exactly how a perfect-looking backtest fails in real trading.
To StratBase's credit, the genuine disclaimers are good. The terms state plainly that past backtest performance does not guarantee future results, that results are simulations that may not reflect real-world conditions, and that the AI does not give financial advice. The vendor's own blog even includes an overfitting warning. StratBase is unregulated software, which is entirely normal for a backtesting tool and not a mark against it. Refunds are a real strength: a “14-day unconditional money-back guarantee on all purchases,” processed through Paddle as Merchant of Record, with cancellation “anytime from your dashboard, no questions asked.” Auto-renewal is disclosed in plain language. The fatal gap is continuity. The terms contain no clause covering what happens to your subscription, your stored strategies, or your data if the business, which is openly for sale, is actually sold or shut down. Data export is allowed within 30 days of termination, then it is deleted.
What traders say
There is effectively no independent voice here, and that absence is the finding. We could not locate a Trustpilot profile, G2, Capterra, Reddit, ForexFactory, or Product Hunt presence. An AlternativeTo listing exists, added in March 2026, with one like and zero reviews. StratBase appears in no third-party backtesting roundup we checked.
The only positive sentiment lives on the vendor's own reviews page: a self-reported 4.9 out of 5 from 12 reviews, every one of them five stars, all posted inside an 18-day window in March 2026, and self-asserted as “verified” without any third-party verification. We anchor our one community-sentiment row to that figure precisely so readers understand what it is: a self-hosted, self-graded number, not an independent rating. For a tool launched in 2024 and asking for monthly payment, the lack of any organic, independent track record is a real concern, not a neutral one.
| Platform | Score | Sample | Source |
|---|---|---|---|
| StratBase.ai (vendor-hosted, self-reported) | 4.9 / 5 | 12 reviews — all 5-star, all within an 18-day March 2026 window; self-asserted “verified,” NOT third-party verified; no independent rating exists | Source ↗ |
| Aggregate | 4.9 / 5 (vendor-hosted) | Self-reported from 12 reviews; no independent rating exists | Normalized by TheFXGeek |
“I've been using StratBase.ai for two weeks and I'm very satisfied. I used to test strategies manually in Excel — it was a nightmare.”
Vendor-hosted review (AndriiSignals), StratBase.ai reviews page (self-hosted, not an independent review), Mar 2026 source ↗
“No independent review of StratBase.ai exists anywhere — no Trustpilot, G2, Capterra, Reddit, or Product Hunt presence. The only ratings are 12 self-hosted, all-5-star reviews posted within an 18-day window, and the business is openly for sale with no continuity clause for subscribers.”
TheFXGeek research note, Independent-coverage check, June 2026, Jun 2026 source ↗
Pros & cons
The balance below is straightforward: StratBase.ai has a modern, capable feature set and honest core disclaimers on one side, and opaque ownership, an active for-sale overhang, oversold data claims, and zero independent validation on the other. The pros are about the product. The cons are mostly about the business behind it.
- Permanent free tier (5 backtests/day, up to 1yr data, all markets)
- Genuine 14-day unconditional money-back guarantee via Paddle (Merchant of Record)
- Named, reputable data vendors: Dukascopy, Binance, Bybit, Alpaca
- AI natural-language strategy builder via Anthropic Claude (shipped)
- 245+ indicators, grid/DCA templates, and multi-timeframe support
- Walk-forward optimisation (single-param on Pro, multi-param on Premium+)
- Honest core disclaimers including a backtest-vs-live caveat and an overfitting warning
- Transparent auto-renewal language and easy “no questions asked” cancellation
- Openly listed for sale at $20,000 with no continuity clause for subscriptions or data if sold or shut down
- No legal entity name, registration number, or named jurisdiction; founder identity unverified
- Forex data is Dukascopy tick aggregated to M1 candles, not “tick-level” as marketed
- Execution modelling (slippage, spread, swap, fills) is contradictory or undocumented
- Self-graded, internally inconsistent accuracy claims supported by likely-fabricated author and review personas
- Web-only with no MT4/MT5/TradingView path, no manual replay, and zero independent reviews
StratBase.ai vs alternatives
Against the obvious free baseline, the MT4 and MT5 built-in strategy testers, StratBase wins on user experience and its AI builder but loses on trust and platform reach. The MetaTrader testers are free, ubiquitous, and the standard place EAs are validated; if your strategy is an EA, you do not need StratBase at all. For discretionary, visual, bar-by-bar practice, FX Replay (which we scored 6.8) is purpose-built and StratBase simply does not compete, because StratBase has no manual replay mode. Against TradingView bar replay, again free to a point, StratBase offers more rigorous automated optimisation but far less charting polish and none of TradingView's enormous community.
Where StratBase is differentiated is the combination of plain-English strategy creation, multi-asset pre-loaded data, and walk-forward optimisation in one browser tab. That is a real niche. The trouble is that established, well-funded competitors do not come with a for-sale banner and an anonymous operator, and for a tool you are trusting to vet real-money strategies, institutional stability matters as much as features.
How StratBase.ai compares to the next tools in our backtesting software ranking:
| Metric | StratBase.ai | FX Replay | Forex Tester |
|---|---|---|---|
| Our score | 5.4/10 | 6.8/10 | 6.8/10 |
| Starting price | Free tier; paid from $29/mo (Pro), Premium $49/mo, Private $99/mo (no lifetime) | Free tier; paid from $17.99/mo or $180/yr (Pro $35/mo, $350/yr) | Free Forexite M1 tier; FTO from €135 one-time; Desktop from $149 + data subscription |
| Best for | AI-assisted, no-code automated backtesting of crypto and US stocks on a free or month-to-month basis | Discretionary forex traders who want deep replay analytics and journaling in the browser | Discretionary traders who want a mature manual bar-replay workflow across many instruments |
| Regulation | Offshore | Offshore | Offshore |
| Full review | This page | FX Replay review → | Forex Tester review → |
In short: FX Replay edges ahead of StratBase.ai in the overall ranking, but aI-assisted, no-code automated backtesting of crypto and US stocks on a free or month-to-month basis is where StratBase.ai makes its strongest case.
Who is StratBase.ai for?
Use StratBase.ai if…
Use StratBase.ai if you want to experiment with AI-assisted, no-code strategy building on crypto or US stocks, you are comfortable staying on the free tier or paying month to month, and you treat every accuracy claim as marketing to be re-tested yourself.
Skip it if…
Skip it if you trade primarily forex and need true tick-level fidelity, you rely on MT4, MT5, or TradingView, you want manual visual replay, or you need confidence that the tool and your saved work will still exist in a year.
Final verdict
StratBase.ai is a promising product attached to a business we cannot ask anyone to depend on. The free tier is worth a look, the refund policy is honest, and the underlying engineering may well be as fast and capable as claimed. But the data is oversold, the execution modelling is half-documented and self-contradictory, the advertised results are self-graded and propped up by unverifiable personas, and the whole thing is openly for sale with no protection for subscribers if it changes hands or disappears. Capable but unproven, opaque, and for sale: that is a tool to sample for free, not to build a real-money validation workflow on. We score StratBase.ai 5.4 out of 10.
A modern, capable cloud backtester with an AI strategy builder and real data vendors, undercut by an active for-sale listing, an anonymous operator, and data marketed as tick-level that is actually aggregated to one-minute candles for forex. Open a demo account to try the platform risk-free, then fund a live account when you're ready. Trading carries risk.
Try StratBase.aiFrequently asked questions
More Backtesting Software reviews
FX ReplayA polished, analytics-rich manual replay tool worth the subscription for discretionary traders, as long as you ignore the “tick-level” marketing and accept web-only delivery plus reliability risk.
Forex TesterA long-established, capable backtester with strong manual replay and honest backtest-vs-live disclaimers, but its premium accuracy rests on tick data from a source the vendor never names.
Traders CasaA clean, low-cost manual backtester that is easy to start with, but you are paying an anonymous operator for a tool that will not disclose its data source, resolution, or fill model.