The engine that reads the markets for you

What the Cerfina AI is

At its core the engine is a statistical system trained to recognise recurring conditions in market data. It continuously scores each monitored instrument for the state it is in, calm, trending or stressed, and for the presence of short-lived patterns that historical analysis has marked as significant. Those micro-trends, small imbalances between buying and selling pressure that appear and fade within hours, are too numerous and too brief for a person to track across a whole market, which is exactly the gap the engine fills.

Its role is deliberately bounded. The AI is a support tool: it prepares information, surfaces conditions and executes within rules you set, but it does not replace your judgement about how much to risk, and it does not form opinions about your life or goals. The platform keeps a human, your personal manager, beside the engine precisely because the tool is powerful, not because it is oracle-like. Anything the engine does is visible, recorded and explained, and anything you disagree with can be stopped.

How the technology works

The pipeline has four stages. First, data collection: prices, volumes and order-book behaviour stream in from connected venues across multiple timeframes. Second, intelligent processing: the raw data is cleaned, aligned and scored, with noise filtered so that meaningful signals stand apart from ordinary jitter. Third, continuous monitoring: every instrument is re-scored as new data arrives, so the picture reflects the market as it is now, not as it was this morning. Fourth, structured presentation: what the engine finds reaches you as short plain-English statements and, where your settings allow, as executed positions with their reasoning attached.

Discipline is built into the mechanics. Exposure scales down when measured volatility rises, because the same nominal position is riskier in a fast market. The volatility circuit breaker can pause trading entirely when conditions cross the thresholds you set. And every decision, executed or skipped, lands in the audit log with its inputs, which is what makes the reports trustworthy rather than merely persuasive.

What the engine analyses

Price movement comes first: levels, momentum and the speed of change across timeframes from minutes to weeks. Trading volume accompanies it, showing where conviction actually sits behind a move. Realised volatility measures how violent recent conditions have been, feeding the sizing and circuit-breaker logic. Trend shape distinguishes orderly movement from exhaustion and reversal patterns. And historical dynamics provide the baseline: how an asset behaved in comparable past regimes, which the engine uses probabilistically rather than deterministically.

Changes in overall market conditions are tracked as a backdrop, because the same price pattern means different things in a calm market than in a stressed one. That context is why the engine can be conservative precisely when less careful systems are busiest.

Advantages of the approach

Speed at which data is processed. The engine evaluates the whole monitored universe continuously. By the time a person has finished reading one chart, it has re-scored thousands of instruments.

Constant monitoring without fatigue. Markets get the same quality of attention at hour eight as at hour one, and at 4 a.m. as at midday, which no human work pattern provides.

Your time back. Users describe the same relief: evenings stop being research shifts, because the watching is handled and summarised.

Information that stays current. The picture in your dashboard is continuously refreshed, so decisions rest on the present state rather than on last night's reading.

Consistency for beginners and experienced users alike. New users get discipline they have not yet built for themselves; experienced users get coverage and consistency across more instruments than they could watch alone.

Who the engine suits

People with limited time who want technology doing the watching. Users of shares and crypto products who want one consistent analytical approach across both. Beginners who want structure and explanations rather than tips. And sceptics, frankly: the engine's decisions are all logged, so a doubtful user can watch it work for a month with minimal exposure and judge the reports against the balances.

It is not for anyone seeking certainty. Probabilities are the engine's language, and they imply being wrong regularly. If a system that sometimes errs by design sounds alarming rather than honest, this is not your tool.

How to use it: four steps

Step 1: registration. Complete the form above and verify your identity. Your manager calls within 24 hours.

Step 2: activation. Fund the account from £200 and configure the engine together with your manager: monitored markets, per-position limits, notification preferences, brake thresholds.

Step 3: getting to know the tools. Spend the first weeks reading: session summaries, the audit history and the volatility dashboard. Ask your manager to explain anything twice if needed.

Step 4: monitoring the market with it. Let the engine run within your settings, review the reports at whatever rhythm suits you, and adjust exposure deliberately rather than emotionally.

A worked example

Consider an illustrative afternoon in a major crypto asset. Overnight the price has drifted up on average volume; at 14:10, volume spikes to several times its recent average while the price accelerates, and measured volatility rises from calm into the top of its normal band. A human watching one chart might read this as breakout excitement and buy. The engine instead does three things in sequence: it scores the pattern against comparable historical episodes, several of which faded within the hour; it reduces the position size it will allow, because volatility has crossed into a faster regime; and it opens a smaller position than usual with a tighter exit, recording the reasoning in the log.

Two hours later the move partially reverses, and the engine exits at its pre-set level with a small loss on a small position. The session report reads plainly: entered on a volume spike, sized down for volatility, exited on reversal, net result slightly negative. Nothing dramatic happened, and that is the point. The interesting comparison is with what excitement alone might have done on that same afternoon with a full-size position and no exit rule. The example is illustrative only, of course: real outcomes vary, and losses can equally accumulate across many such sessions.

Questions people ask about the engine

See the engine work

Register free, talk through the settings with your manager, and watch the engine on a small balance for a month. The reports will teach you more about its behaviour than this page can, and every conclusion will be your own.