Modules Guide

Every module — what it does and how to use it

Practical guidance on each module in Dome Terminal — ORACLE, the AI Committee, Bots, Journal, Alerts, and Academy. Written for traders who want to use the tools well, not just know they exist.

ORACLE AI — Your context-aware analyst

ORACLE is an AI assistant that's already read your chart, your risk state, and your recent journal before you ask your first question. It's not a generic chatbot — every answer comes from what's actually in your terminal right now.

What ORACLE already knows when you open it

Your active chart symbol, timeframe, and current price action. Quant Brain readings — trend, volatility, structure, delta. Your last 30 journal entries and any mistake tags. Current portfolio positions and PnL. Your risk settings and daily loss status. The more complete your chart and journal are, the more specific the answers you'll get.

How to get the most out of ORACLE

01

Open from the sidebar

Click the ORACLE icon in the left sidebar or press Alt+O. The panel opens alongside your chart — keep it docked or float it to a second monitor. It retains the conversation history for the current session.

02

Have your chart loaded first

Before you ask a market question, make sure your chart is loaded and Quant Brain is active. ORACLE reads the current state — if your workspace is empty, it has nothing to work with, and you'll get generic answers.

03

Ask specific questions

Vague questions get vague answers. "Will BTC go up?" gives ORACLE nothing useful to work with — there's no context in the question, so the answer can't be specific either. The prompt comparison below shows what good questions actually look like.

Useful prompts
"Summarize the 4H market state for BTC using only what you see in the terminal. Is the structure consistent with a continuation long from this level?"
"My journal shows I keep exiting too early on trend days. Based on my last 10 entries, what pattern do you see?"
"Quant Brain shows high volatility and bearish structure on the 1H. Does my current long position size still make sense given those conditions?"
Specific, tied to what ORACLE can actually see, and asks for reasoned analysis — not a prediction.
Prompts that don't work
"What's the best trade right now?"
"Will Bitcoin go up or down today?"
"Should I buy or sell?"
These give ORACLE nothing to work with. There's no useful context in the question, so there can't be a useful answer. ORACLE can only tell you what it sees — not what will happen.

What ORACLE can't do — and why that's intentional

ORACLE can't trade for you, change your settings, or make decisions. It's an analyst, not an autopilot. Every trade decision, entry, and action stays with you. That's by design — the goal is to make you a better decision-maker, not to replace your judgment. Every ORACLE session is logged automatically and can be reviewed from Journal → ORACLE History.

AI Committee — Eight angles on one trade

When you want a proper pressure-test on a trade idea, the AI Committee runs 8 specialist AI agents against it simultaneously. Each one comes from a different angle. They can disagree — and that disagreement is useful information.

Submit a trade idea — symbol, direction, and your thesis in a few sentences — and all 8 specialists analyze it in parallel. Pay closest attention to any specialist showing a Hard Block. That means conditions that historically correlate with poor outcomes in that specialist's area. It's not a suggestion — it's a warning.

📈
Technical
Price structure & pattern
▲ Favorable
Evaluates price structure, chart pattern quality, key levels, and whether the proposed entry aligns with technical context.
Orderflow
Delta & depth analysis
→ Neutral
Analyzes cumulative delta, absorption events, bid/ask imbalances, and liquidity depth relative to the proposed direction.
📡
Sentiment
News & social signals
▲ Favorable
Reads market news, social signal data, and narrative momentum to evaluate whether broader sentiment supports the trade.
🛡
Risk
Sizing & drawdown exposure
⊘ Hard Block
Evaluates proposed position size against your account size, current drawdown, daily loss gate status, and correlation with existing positions.
🔗
On-Chain
Blockchain metrics (crypto)
▲ Favorable
For crypto assets: evaluates on-chain flows, exchange inflows/outflows, whale activity, and network health metrics.
🌐
Macro
Economic context
→ Neutral
Evaluates the macroeconomic backdrop: rate environment, risk-on/off regime, recent economic data, and central bank posture.
📊
Quant
Statistical edge
▲ Favorable
Reviews the statistical history of similar setups: base rate, average reward-to-risk, win rate in the current regime, and OOS performance.
Entry Timing
Execution precision
⚠ Wait
Evaluates whether right now is the optimal entry moment: spread conditions, time-of-day patterns, and whether the setup has fully developed.

Pay most attention to Hard Blocks

A Hard Block from the Risk specialist means the position size or current drawdown state makes the trade structurally unsound — regardless of how good the chart looks. Hard Blocks are not suggestions. They reflect conditions associated with poor outcomes. You can override them, but you should have a clear, documented reason for doing so.

Bot Management — Automated strategy execution

Bots execute Dome Script strategies automatically according to your configured conditions. They don't replace your judgment — they apply rules you've already defined, exactly as written, without emotional interference. The key word is "already defined" — bots are only as good as the strategy and risk gates you give them.

01

Create from a Script Library strategy

Navigate to Script Library, open a validated strategy, and click Create Bot. You cannot create a bot from an unvalidated script — the validation step is mandatory. This ensures the strategy logic runs without errors before any capital is involved.

02

Configure risk gates

Before anything else, set the three risk gates (described below). Without risk gates, a bot running on a broken strategy or in adverse market conditions has no automatic stopping mechanism. Risk gates are non-optional for any real-money deployment.

03

Enable paper mode — run for at least 2 weeks

Paper mode simulates execution against live market data without placing real orders. Run in paper mode long enough to cover a meaningful variety of market conditions — at minimum two calendar weeks, ideally a full month. Do not skip this step.

04

Review the bot logs critically

After the paper period, open Bot → Logs and review every execution. Look for: fills that shouldn't have triggered, missed signals, unusual clustering of losses, and whether performance matches what the backtest predicted. Unexplained discrepancies are a warning sign — investigate before going live.

05

Consider live mode only after thorough log review

Switching to live mode should be a deliberate decision based on evidence from the paper log — not impatience. Even after switching, start with reduced position sizing and expand gradually as the live performance builds a track record.

The three risk gates

Daily Loss Gate
Sets a maximum loss the bot can accumulate in a single calendar day. When reached, the bot stops all new executions until midnight. Example: if your daily gate is 1.5% of account and losses hit that level at 11am, the bot pauses until the next session. This prevents a single bad day from compounding into account-threatening drawdown.
Maximum Position Gate
Caps the total notional value any single position can reach, regardless of what the strategy signals. This prevents a runaway entry signal from building an oversized position. Set this to a value you would be genuinely comfortable with in a worst-case scenario — not optimistic about.
Correlation Gate
Prevents the bot from opening new positions that are highly correlated with existing open positions. If you're already long ES and the bot wants to go long NQ on the same signal type, the correlation gate can block it — because two correlated positions in the same direction is effectively one larger, riskier position.
Trade Journal — Behavioral evidence

Your journal is the most underused tool in the terminal for most traders. It converts trading from a series of disconnected events into a structured dataset about your behavior — one that ORACLE can analyze for patterns you won't notice yourself.

What fields matter and why

Setup
Tag the specific setup type (e.g., "breakout retest", "failed auction", "trend continuation"). This is the field ORACLE uses to calculate which setups are performing and which are degrading.
Planned vs. Taken
Did you take the trade as planned, or did you deviate (size up, move your stop, enter early)? This single field, reviewed monthly, reveals most behavioral patterns.
Emotional state
Calm, anxious, frustrated, overconfident? Rate it 1–5. Correlating your emotional state with trade outcomes over 50+ entries is consistently revealing.
Mistake tag
If you broke a rule, tag it specifically: "sized up outside plan", "chased entry", "moved stop for no reason". Vague tags ("bad day") are useless for analysis.
Market regime
What was Quant Brain showing? Trending or ranging? High or low volatility? This links your result to the market context automatically.
Lesson
One sentence: what would you do differently? If you have nothing to write here, the entry isn't complete. Even on winning trades, there's usually something to note.

AI-assisted journal review

After 20+ journal entries, open ORACLE and ask: "Review my last 30 journal entries. What patterns do you see in my mistake tags and emotional state ratings?" ORACLE will surface clusters you may have missed — like "you lose more on Mondays" or "your biggest losses come when emotional state was rated 4 or 5". Use this monthly as a behavioral audit.

Smart Alerts — Reduce screen time

Smart Alerts let you step away from the screen and come back when something meaningful has happened — not just when price moved a random amount. The goal is fewer, higher-quality notifications, not a constant stream of pings.

Alert types available

Price Level Alert
Quant Brain State Alert
Portfolio Risk Alert
Review Reminder
Alert typeWhat triggers itBest use case
Price LevelPrice touches or crosses a level you define on the chartDraw a level, right-click → Set Alert. Better than watching the chart manually for a level that may or may not be tested that session.
Quant Brain StateA Quant Brain metric changes state (e.g., Volatility moves from Normal to High, or Trend changes direction)Alert on regime changes — let Quant Brain watch the chart while you're away, and notify you when the context shifts meaningfully.
Portfolio RiskYour total portfolio risk exposure reaches a threshold you setUseful when running multiple bots or positions — get alerted before you breach your overall risk limit, not after.
Review ReminderTime-based — fires at a scheduled time you configureSet a daily 5pm reminder to log today's trades in the journal before you close the terminal.

Alert chains — connecting alerts to bot actions

You can chain a Portfolio Risk Alert to a bot pause action: when your portfolio loss exceeds X%, the alert fires and simultaneously pauses all active bots. Configure this under Alerts → Actions → Link to Bot. This is a useful safety mechanism that doesn't require you to be watching the screen during a drawdown event.

Academy — Structured learning

Academy is not a generic trading course library. It's built around the tools in Dome Terminal and the behavioral patterns most commonly found in traders' journals. Lessons connect directly to the features you're using.

Learning paths by trader type

🎯
Discretionary Trader
Chart reading, Quant Brain interpretation, journal discipline, multi-timeframe framework, and behavioral review cycles.
⚙️
Systematic Trader
Dome Script fundamentals, backtest methodology, walk-forward validation, bot risk gate configuration, and log analysis.
🏢
Prop Firm Trader
Risk rule integration, daily loss gate setup, drawdown management, journal requirements, and challenge-phase discipline.

How Mentor mode works

Mentor mode in Academy observes your terminal activity (with permission) and surfaces relevant lesson suggestions based on what you're doing. If you're reviewing a losing trade and you tagged "moved stop impulsively" in the journal, Mentor mode may surface the lesson on stop management discipline. It's reactive guidance, not a lecture series.

Enable Mentor mode from Academy → Settings → Mentor. You can disable it anytime and it logs no data externally.

Connect Academy lessons to your journal mistakes

After a monthly journal review with ORACLE, you'll have a list of recurring mistake tags. Use that list to search Academy for lessons that address those specific patterns. Learning is most effective when it's directed at a mistake you've already identified in your own data — not at a topic chosen at random.