18 modules. One composite score.
Every module watches one dimension of risk in real time and feeds a single AI-weighted score for each account. All modules are included in every plan.
HFT & cheater detection
Trades-per-second analysis, burst detection and profit-concentration scoring, rolled into a live cheater score from 0 to 100.
Abuse & toxic flowIP risk monitor
Flags VPN, proxy and TOR use, geographic jumps and accounts sharing IPs or devices, each with its own risk weight.
Abuse & toxic flowMargin & leverage monitor
Live margin level and effective leverage per account, with early warnings before a margin call lands.
Exposure & marginBehavioral risk monitor
Spots revenge trading, overtrading and tilt, and builds an emotional-risk score per account with thresholds you set.
Behavior & strategyConcentration risk
Exposure concentration by symbol, client and group, with diversity scoring and net position tracking.
Exposure & marginScalper monitor
Trade-frequency and hold-time distributions that classify scalping intensity over configurable windows.
Abuse & toxic flowMartingale & grid detection
Finds lot-size escalation after losses, recovery sequences and grid ladders using price-level variance.
Behavior & strategyTrading strategy monitor
Combines pattern detectors to classify each account's strategy and publishes daily strategy statistics.
Behavior & strategySwap manager
Tracks position age and swap-avoidance behaviour and measures the revenue impact by trader and group.
Exposure & marginFast deals monitor
Isolates ultra-short trades under 60 seconds across five duration buckets and highlights suspicious clusters.
Abuse & toxic flowToxic flow detection
Scores every account on profitability against your book, so you can route informed flow to A-book with evidence.
Abuse & toxic flowLatency arbitrage detector
Compares fill times with reference feeds to expose traders consistently trading on stale prices.
Abuse & toxic flowNews event guard
Watches order bursts around scheduled economic releases and applies stricter rules inside news windows.
Behavior & strategyBonus abuse detector
Links accounts with mirrored or opposite positions during promotions and flags hedged bonus farming.
Abuse & toxic flowDrawdown manager
Daily and overall drawdown limits per account or group, built for prop-firm challenge rules.
Exposure & marginNet exposure & hedging
Real-time net exposure by symbol across the whole book, with suggested hedge sizes for your LP.
Exposure & marginAlerts & reporting
Routes alerts to dashboard, email or Telegram and generates scheduled risk reports for management and audit.
AI & operationsAI risk copilot
Ask questions in plain English and get explained answers: why an account was flagged and what to do next.
AI & operationsHFT & cheater detection
Trades-per-second analysis, burst detection and profit-concentration scoring, rolled into a live cheater score from 0 to 100.
Why it matters
High-frequency abuse is the fastest way for a single account to pull money out of a B-book. Bots and EAs fire dozens of orders a second, exploit tiny price gaps and close before the market moves. By the time a daily report shows the profit, it has already been withdrawn.
What it detects
How it works
- 1
Every deal is scored the moment it is executed, using rolling windows of 1 s, 10 s, 1 min and 1 h.
- 2
Signals are weighted by the AI model against the account's own history and its group's normal profile.
- 3
The result is a continuous cheater score from 0 to 100, updated live, with the top reasons attached.
Key metrics
- Cheater score 0–100
- Trades per second (peak and average)
- Share of profit from trades under 10 s
- Burst count per session
- Average and median hold time
- Score trend over 24 h / 7 d
Actions your desk can take
- Move the account to A-book or a dedicated group
- Apply a minimum hold-time rule
- Freeze trading pending review
- Open a compliance case with the evidence attached
Configurable settings
- Trades-per-second thresholds per group
- Score weights for each signal
- Alert levels (watch, warning, critical)
- Whitelist for approved algo clients
Will legitimate algo traders be flagged?
Approved algorithmic clients can be whitelisted or placed in a group with their own thresholds, so only abusive behaviour is scored as high risk.
How quickly is a burst detected?
Bursts are recognised as the deals arrive. Alerts typically reach the desk in under 100 milliseconds.
IP risk monitor
Flags VPN, proxy and TOR use, geographic jumps and accounts sharing IPs or devices, each with its own risk weight.
Why it matters
One person running ten accounts can farm bonuses, hedge against themselves or bypass leverage limits. IP and device intelligence is the quickest way to connect accounts that look unrelated on paper.
What it detects
How it works
- 1
Login and connection events are captured from MT5 or cTrader with IP and terminal data.
- 2
Each IP is enriched with network type, location and reputation, then compared with every other account.
- 3
Accounts are grouped into clusters, and each account receives an IP risk score with its linked accounts listed.
Key metrics
- IP risk score 0–100
- Linked account clusters
- VPN / proxy / TOR flags
- Countries per account
- First and last seen per IP
Actions your desk can take
- Review the linked cluster in one view
- Block bonus eligibility for linked accounts
- Request additional KYC
- Restrict logins from specific networks
Configurable settings
- Cluster time window
- Restricted country list
- Allowed corporate IP ranges
- Score weights per signal
Does it need extra software on the client side?
No. It uses the login and connection data your platform already records.
Can families or offices sharing an IP be excluded?
Yes. You can mark known shared IPs or ranges as approved so they don't raise the score.
Margin & leverage monitor
Live margin level and effective leverage per account, with early warnings before a margin call lands.
Why it matters
Margin problems escalate quickly in volatile markets. Without live visibility, brokers discover negative balances and forced stop-outs after the fact, along with the client complaints that follow.
What it detects
How it works
- 1
Positions and quotes stream continuously, so equity and margin are recalculated on every tick that matters.
- 2
A volatility model estimates the price move needed to reach margin call and the time it may take.
- 3
Accounts are ranked by urgency and sent to the desk with the required deposit already calculated.
Key metrics
- Margin level %
- Effective leverage
- Distance to margin call / stop-out
- Required deposit amount
- Accounts at risk by group
Actions your desk can take
- Send an automated margin warning to the client
- Notify the account manager or IB
- Reduce leverage for the account or group
- Prepare for stop-out ahead of news
Configurable settings
- Warning levels per group
- Volatility source and window
- Regulatory leverage caps by jurisdiction
- Notification channels
Can it notify clients directly?
Alerts can be pushed to your CRM or email system through webhooks so clients and account managers are warned early.
Does it support different leverage rules per region?
Yes. Caps can be set per jurisdiction, group or account type.
Behavioral risk monitor
Spots revenge trading, overtrading and tilt, and builds an emotional-risk score per account with thresholds you set.
Why it matters
Emotional trading changes how clients behave and how much risk they bring. Revenge trading after a loss, overtrading and sudden size changes are early signs of accounts that will either blow up or turn unexpectedly toxic.
What it detects
How it works
- 1
The engine builds a behavioural baseline for every account from its own trading history.
- 2
New activity is compared with that baseline and with the group profile in real time.
- 3
An emotional-risk score is produced with the behaviours driving it listed in plain language.
Key metrics
- Behavioural risk score 0–100
- Revenge-trade count
- Overtrading ratio vs. baseline
- Lot-size escalation factor
- Session tilt markers
Actions your desk can take
- Alert the account manager for a client call
- Apply temporary leverage reduction
- Flag for responsible-trading review
- Adjust B-book exposure limits for the account
Configurable settings
- Baseline period
- Threshold per behaviour
- Score weights
- Groups to include or exclude
Is this used against clients?
It is a risk and client-protection tool. Many brokers use it to support responsible-trading policies and early contact with clients.
How much history is needed?
A baseline starts after a few days of trading and becomes more accurate over time.
Concentration risk
Exposure concentration by symbol, client and group, with diversity scoring and net position tracking.
Why it matters
A book can look balanced overall while a single symbol, client or IB group carries most of the risk. Concentration is often what turns a normal market move into a painful day.
What it detects
How it works
- 1
Open positions are aggregated live by symbol, client, group and IB.
- 2
Concentration and diversity indices are calculated and compared with your limits.
- 3
Breaches and fast-growing concentrations trigger alerts with the largest contributors listed.
Key metrics
- Top-10 client share of exposure
- Exposure by symbol and direction
- Diversity score
- Correlated exposure groups
- Limit utilisation %
Actions your desk can take
- Hedge the concentrated symbol with your LP
- Move top contributors to A-book
- Set per-symbol exposure caps
- Alert dealing before news events
Configurable settings
- Concentration limits per symbol and group
- Correlation groups
- Alert thresholds
- Reporting currency
Does it include correlation between symbols?
Yes. You can define correlated groups, and the engine also suggests correlations from price history.
Can I see it by IB?
Yes. Exposure can be grouped by IB, group, account type or any custom tag.
Scalper monitor
Trade-frequency and hold-time distributions that classify scalping intensity over configurable windows.
Why it matters
Not all scalping is a problem, but aggressive scalping on tight spreads can capture value from the broker with almost no market risk. The key is separating normal short-term traders from those systematically exploiting spread and execution.
What it detects
How it works
- 1
Each closed trade is placed into hold-time buckets and compared with configurable windows.
- 2
Profit per bucket and spread capture are calculated per account.
- 3
A scalping risk score separates healthy short-term trading from exploitative patterns.
Key metrics
- Scalping risk score 0–100
- Trades per hour
- Share of trades under 1 min / 5 min
- Profit per hold-time bucket
- Spread capture ratio
Actions your desk can take
- Apply minimum hold-time rules
- Move to a raw-spread or A-book group
- Adjust execution settings for the group
- Flag for dealing review
Configurable settings
- Time windows
- Bucket boundaries
- Thresholds per group
- Symbols to include or exclude
Is scalping always flagged?
No. Scoring focuses on whether short trades are consistently extracting value, not on the trading style itself.
Can windows be changed per group?
Yes. ECN and standard groups can use different windows and thresholds.
Martingale & grid detection
Finds lot-size escalation after losses, recovery sequences and grid ladders using price-level variance.
Why it matters
Martingale and grid strategies look profitable for long periods, then fail all at once. When they do, the result is often a margin cascade, negative balances and sudden exposure for the broker.
What it detects
How it works
- 1
Trade sequences are reconstructed per account and symbol.
- 2
Volume progression and price spacing are analysed to recognise martingale and grid structures.
- 3
A strategy risk score is produced with the estimated blow-up price level.
Key metrics
- Martingale / grid score 0–100
- Max lot multiplier in sequence
- Grid spacing and depth
- Floating loss on basket
- Estimated stop-out price
Actions your desk can take
- Warn the client before the sequence deepens
- Cap maximum position size
- Hedge the basket exposure
- Move to a monitored group
Configurable settings
- Multiplier threshold
- Minimum sequence length
- Grid spacing tolerance
- Alert levels
Can it tell grid from normal scaling in?
Yes. It looks at regular price spacing and volume progression together, which normal scaling-in rarely shows.
Does it estimate risk to the broker?
It estimates the price at which the basket would hit stop-out, so dealing can plan hedges ahead.
Trading strategy monitor
Combines pattern detectors to classify each account's strategy and publishes daily strategy statistics.
Why it matters
Knowing how each client trades lets you price, route and hedge them correctly. The strategy monitor gives every account a clear classification instead of a long list of separate flags.
What it detects
How it works
- 1
Outputs from all pattern detectors are combined per account.
- 2
The AI model assigns a primary and secondary strategy with a confidence level.
- 3
Daily statistics and changes are published to the dashboard and reports.
Key metrics
- Primary / secondary strategy
- Confidence %
- Daily strategy statistics
- Strategy change alerts
- P&L by strategy type
Actions your desk can take
- Route strategy types to A-book or B-book
- Price groups by strategy
- Target marketing at profitable segments
- Review sudden style changes
Configurable settings
- Strategy definitions
- Minimum trades for classification
- Change sensitivity
- Report schedule
Can we add our own strategy types?
Yes. Custom strategy definitions can be added during onboarding or through enterprise customisation.
How often are classifications updated?
Continuously, with a daily summary snapshot for reporting.
Swap manager
Tracks position age and swap-avoidance behaviour and measures the revenue impact by trader and group.
Why it matters
Swap is a quiet revenue line that some traders learn to avoid, by closing before rollover and reopening after, or by using swap-free accounts for carry strategies. Over time this adds up to real lost income.
What it detects
How it works
- 1
Positions are monitored around rollover times for every symbol.
- 2
Patterns of closing and reopening, and swap-free carry, are recognised automatically.
- 3
Revenue impact is estimated and traders are classified by swap risk.
Key metrics
- Swap avoidance score
- Estimated lost swap revenue
- Average position age
- Swap-free accounts at risk
- Triple-swap activity
Actions your desk can take
- Review swap-free eligibility
- Apply admin fees per your policy
- Adjust swap settings for the group
- Report revenue impact to management
Configurable settings
- Rollover window
- Swap-free rules
- Revenue currency
- Alert thresholds
Does it support Islamic accounts?
Yes. It is designed to help enforce fair use of swap-free accounts under your own policy.
Is lost revenue estimated in money terms?
Yes, in your reporting currency, by account, group and period.
Fast deals monitor
Isolates ultra-short trades under 60 seconds across five duration buckets and highlights suspicious clusters.
Why it matters
Ultra-short trades are where most execution abuse hides. Looking at them separately, across clear duration bands, makes suspicious patterns far easier to see.
What it detects
How it works
- 1
Closed trades under 60 seconds are captured and placed into the five buckets.
- 2
Win rate and profit are calculated per bucket and compared with peers.
- 3
Accounts with suspicious concentration are scored and escalated.
Key metrics
- Fast-deal score 0–100
- Trades per bucket
- Profit and win rate per bucket
- Top symbols and sessions
- Peer percentile
Actions your desk can take
- Investigate execution on targeted symbols
- Apply hold-time rules
- Move to A-book
- Escalate to HFT review
Configurable settings
- Bucket boundaries
- Minimum trade count
- Peer group definition
- Alert levels
Why five buckets?
They separate the very fastest trades, where abuse is most likely, from normal short-term trading.
Can bucket ranges be changed?
Yes. Bucket boundaries are fully configurable.
Toxic flow detection
Scores every account on profitability against your book, so you can route informed flow to A-book with evidence.
Why it matters
Toxic flow comes from traders who are consistently right against your book, whether through skill, information or execution edge. Keeping it in the B-book turns market-making into a steady loss.
What it detects
How it works
- 1
Every trade is marked out against subsequent prices at several horizons.
- 2
Mark-outs, profitability and edge signals are combined into a toxicity score.
- 3
The AI suggests A-book, B-book or hybrid routing with the evidence behind it.
Key metrics
- Toxicity score 0–100
- Mark-out P&L at 1 s / 5 s / 30 s / 60 s
- Profit against book
- Suggested routing
- Toxicity trend
Actions your desk can take
- Move account to A-book
- Use hybrid routing by symbol
- Adjust spread or execution group
- Review with dealing and compliance
Configurable settings
- Mark-out horizons
- Minimum trade sample
- Routing thresholds
- Symbols in scope
Is a profitable trader always toxic?
No. Toxicity looks at how and why the trader wins. Skilled swing traders can be handled differently from latency or information-driven flow.
Can it connect to our bridge?
Routing suggestions can be pushed by webhook or API to your bridge or dealing tools, including Rnix Liquidity Bridge.
Latency arbitrage detector
Compares fill times with reference feeds to expose traders consistently trading on stale prices.
Why it matters
Latency arbitrage uses a faster price feed to trade on quotes that are already out of date. Each trade is small, but done at scale it drains spread revenue every day.
What it detects
How it works
- 1
Fills are compared with reference prices from your LP or market data source.
- 2
Price lead, reaction time and exit timing are calculated per trade.
- 3
A latency-arbitrage score is built per account, with the worst trades listed as evidence.
Key metrics
- Latency-arb score 0–100
- Average price lead (ms)
- Share of fills on stale quotes
- Profit from stale fills
- Most targeted symbols
Actions your desk can take
- Move to A-book immediately
- Adjust price feed or execution delay for the group
- Cancel trades per your terms
- Escalate to compliance
Configurable settings
- Reference feed source
- Lead-time threshold
- Minimum trade sample
- Symbols in scope
What reference feed is used?
You can use your LP prices, bridge prices or another market data source agreed during onboarding.
Does it work for crypto CFDs?
Yes, as long as a reference feed is available for the symbol.
News event guard
Watches order bursts around scheduled economic releases and applies stricter rules inside news windows.
Why it matters
Major data releases create the biggest price gaps and the most dangerous execution conditions. Bots that trade the first milliseconds of a release can do a lot of damage in a few seconds.
What it detects
How it works
- 1
Scheduled events are loaded from the economic calendar with their affected symbols.
- 2
Activity in configurable windows around each release is analysed separately.
- 3
Accounts and exposure at risk are flagged before and during the event.
Key metrics
- News-trading score
- Orders inside release windows
- Profit from release gaps
- Exposure on affected symbols
- Upcoming high-impact events
Actions your desk can take
- Apply stricter rules during news windows
- Hedge exposure before the release
- Restrict pending orders near events
- Review repeat news traders
Configurable settings
- Event impact levels
- Window before / after release
- Affected symbols per event
- Alert timing
Can it apply different rules during news?
Yes. Thresholds can tighten automatically within the release windows you define.
Does it warn the desk ahead of events?
Yes. You get a pre-event exposure summary for affected symbols.
Bonus abuse detector
Links accounts with mirrored or opposite positions during promotions and flags hedged bonus farming.
Why it matters
Promotions are meant to win clients, not pay arbitrageurs. Hedged bonus trading across linked accounts can turn an entire campaign budget into withdrawals.
What it detects
How it works
- 1
Accounts are linked using IP, device and behavioural matching.
- 2
Positions across linked accounts are compared for hedging and mirroring.
- 3
Each promotion gets an abuse score and a list of suspect pairs or groups.
Key metrics
- Bonus abuse score 0–100
- Linked account pairs
- Hedged volume
- Bonus at risk
- Abuse rate per campaign
Actions your desk can take
- Freeze or remove bonus per your terms
- Block linked accounts from future campaigns
- Request extra verification
- Report campaign ROI to marketing
Configurable settings
- Promotion groups
- Link signals and weights
- Hedging time tolerance
- Alert levels
Can CRM data be used?
Yes. Payment and KYC data from your CRM, such as Rnix Core CRM, can strengthen account linking.
Does it help marketing?
It shows the real cost and abuse rate of each promotion, so budgets go to campaigns that work.
Drawdown manager
Daily and overall drawdown limits per account or group, built for prop-firm challenge rules.
Why it matters
For prop firms, drawdown rules are the product. For brokers, drawdown on managed or PAMM accounts is a key client-protection metric. Both need to be enforced in real time, not reconciled the next morning.
What it detects
How it works
- 1
Equity is tracked live against each account's starting and peak balance.
- 2
Rule sets are applied per challenge, phase or group.
- 3
Breaches trigger alerts or automatic actions, with a full audit trail.
Key metrics
- Daily and max drawdown %
- Distance to breach
- Rule violation log
- Profit target progress
- Pass / fail status
Actions your desk can take
- Close positions or disable trading on breach
- Notify trader and support team
- Move to next challenge phase
- Export results for payout review
Configurable settings
- Rule sets per program
- Static, trailing or equity-based limits
- Automatic action on breach
- Reset time and time zone
Can it run multiple prop programs?
Yes. Each challenge or funding program can have its own rule set.
Are automatic actions optional?
Yes. You can choose alert-only, manual approval or fully automatic actions.
Net exposure & hedging
Real-time net exposure by symbol across the whole book, with suggested hedge sizes for your LP.
Why it matters
Net exposure is the broker's real market risk. Seeing it live, by symbol and currency, is what lets the dealing desk hedge at the right time and size.
What it detects
How it works
- 1
All open positions are aggregated live by symbol, currency and book.
- 2
Exposure is compared with limits and with recent volatility.
- 3
Hedge sizes are suggested for your LP and breaches are alerted instantly.
Key metrics
- Net exposure by symbol / currency
- Broker floating P&L
- Limit utilisation %
- Suggested hedge volume
- Value-at-risk estimate
Actions your desk can take
- Hedge with your liquidity provider
- Move contributors to A-book
- Tighten limits before events
- Share exposure reports with management
Configurable settings
- Exposure limits per symbol
- Base / reporting currency
- VaR horizon and confidence
- Hedge rounding rules
Can it place hedges automatically?
Hedge suggestions can be sent through API or webhook to your bridge. Automatic execution can be configured in enterprise setups.
Does it include VaR?
Yes. A value-at-risk estimate is available with configurable horizon and confidence.
Alerts & reporting
Routes alerts to dashboard, email or Telegram and generates scheduled risk reports for management and audit.
Why it matters
Detection only helps if the right person sees it at the right moment, and if management and compliance can review what happened later.
What it detects
How it works
- 1
Each module publishes alerts with severity, account, reasons and evidence.
- 2
Routing rules send them to the right people and channels.
- 3
Reports are generated automatically on your schedule and stored for audit.
Key metrics
- Alerts by module and severity
- Response time per alert
- Open vs. resolved cases
- Daily risk summary
- Regulatory exports
Actions your desk can take
- Acknowledge and assign alerts
- Open and track cases
- Schedule management reports
- Export evidence for compliance
Configurable settings
- Routing rules
- Quiet hours and escalation
- Report schedules and recipients
- User roles and permissions
Which channels are supported?
Dashboard, email, Telegram, webhooks and REST API. Slack or Teams can be added for enterprise plans.
Can reports be white-labelled?
Yes. Reports can carry your company branding.
AI risk copilot
Ask questions in plain English and get explained answers: why an account was flagged and what to do next.
Why it matters
Risk teams spend hours pulling data together to answer simple questions. The copilot lets anyone on the desk ask in plain English and get an answer, with the trades and accounts behind it.
What it detects
How it works
- 1
The copilot reads the same live data and scores as the rest of RNIX AI Desk.
- 2
It answers questions using your book's data, with the supporting records attached.
- 3
It recommends actions, and your team decides. Nothing is executed without permission.
Key metrics
- Answers with linked trades and accounts
- Daily briefings
- Case summaries
- What-if estimates
- Suggested next actions
Actions your desk can take
- Ask which accounts drove today's losses
- Summarise a case for compliance
- Estimate the effect of A-book routing
- Prepare a management briefing
Configurable settings
- Users allowed to use the copilot
- Briefing schedule
- Data scope per role
- Language preferences
Does our data train other models?
No. The copilot runs within your deployment and your data isn't used for other clients.
Can it take actions on its own?
No. It recommends actions. Any change requires a user with the right permission.
Put an AI risk desk on every trade.
See RNIX AI Desk running against your own MT5 or cTrader data in a live walkthrough with our team.