About eight months ago, I started learning to read charts. Not as a career move — just curiosity. I was reading about how markets move, watching YouTube videos about ICT concepts (liquidity, order blocks, fair value gaps), and somewhere in that rabbit hole I got hooked.
The problem: I work full-time as a Lead DevOps Engineer at DigitalOcean. I can't sit and watch charts all day. And the best trading sessions for XAU/USD — the London open, the New York killzone — sometimes happen at 2 AM my time in Karachi.
So I did what any DevOps engineer with too much free time would do: I built a system.
"You don't advise. You act." — MORPHEUS system prompt
What Is MORPHEUS?
MORPHEUS is my personal AI operator. It's not a product. It's not open source (yet). It runs entirely on my EC2 instance via Docker Compose and connects to my live trading account, Kubernetes clusters, Slack workspace at work, GitHub, and my local machine — all at the same time.
Think of it as a best friend who happens to also manage your entire tech stack and trade for you when you're asleep.
The frontend is a React + TypeScript app with a Three.js particle orb that pulses when MORPHEUS is speaking. The backend is Node.js + Express with Socket.IO for real-time streaming. All intelligence comes from OpenAI GPT-4o with tool calling — 73 tools registered across 14 categories.
The Trading Side
This is the part people find hard to believe. MORPHEUS has full, live integration with my MT5 trading account — it can check my balance, see open positions, place and close trades, and automatically manage risk. All from a chat interface or by voice.
There are multiple bots running, each focused on a different session or approach. They run on a schedule, scan the market, and either find a setup or skip — without me needing to watch a screen.
Asian Session Order Block Bot
This bot runs during a specific trading session and looks for a particular price pattern on the H1 chart. When conditions are met, it enters automatically — and if nothing qualifies, it simply doesn't trade. It's the most backtested part of the system with strong historical results.
Gold Scalper — Multi-Pair Watchlist
A scalping bot that continuously monitors 14 pairs — metals, forex, indices, and crypto. What makes it different is the AI gate: before any trade is placed, GPT-4o-mini reviews the setup and gives a confidence score. Only setups above the threshold go through. This single addition made the biggest difference in quality of entries.
Chart Analyser
I can describe what I'm looking for in plain English and MORPHEUS pulls live price data, renders an annotated chart, and gives me a structured trade plan. No switching tabs, no manual drawing — just ask and it appears.
Top-Down Analysis
Before a trading session I ask for a full top-down read. MORPHEUS goes through multiple timeframes, identifies the overall bias, key levels, and hands me a final entry plan. Something that used to take 30 minutes now takes about 15 seconds.
MT5 Live Trading Panel
A dedicated panel showing the live account at all times — balance, equity, margin, every open position, and a breakdown of how MORPHEUS-tagged trades are performing. Trades can be placed, closed, or modified from here or directly via chat.
Trade Setup Scanner
An automated scanner that runs across multiple pairs and surfaces trade setups as cards — with entry, SL, TP, and direct execution buttons. Results are tracked with P&L history so I can see how each setup type performs over time.
Trading Journal
Every trade — whether placed manually, by the Gold Scalper, OB Bot, or any other bot — flows into a single unified journal. Full history, P&L tracking, and session tags in one view.
Session Reclaim Bot
A bot focused on session transition moments — when the market sweeps a key level then aggressively reverses back. It runs automatically and only trades when the pattern is clean.
Session IFVG Bot
Auto-starts at a scheduled time each day, targets key previous day levels, and looks for a specific entry pattern. Fully hands-off — it either finds a trade or it doesn't.
H1 OB → M1 FVG Bot
A multi-timeframe bot that uses a higher timeframe to identify structure and a lower timeframe for a precision entry trigger. Combines two timeframes to improve entry accuracy.
Pyramid Scalper
A scaling bot that adds to a winning position as it moves in favour. Available for both XAU/USD and BTC, each with separate configurations suited to the instrument's volatility.
AI Asset Analyzer
On-demand deep analysis of any asset. MORPHEUS pulls live data and produces a structured bias report — trend direction, key levels, and a trade recommendation across timeframes.
Kronos Scalper
Combines the Kronos forecasting model with scalping logic to produce a structured trade recommendation — direction, confidence score, entry, SL, and TP — on the M5 timeframe.
DOCN Stock Monitor
A small personal touch — a dedicated panel monitoring DigitalOcean (DOCN) stock. Since it's my employer, I like keeping an eye on it alongside everything else.
The DevOps Side
I built MORPHEUS as a trader's assistant first — but I'm a DevOps engineer full-time, so naturally it became a DevOps tool too.
From the MORPHEUS chat, I can:
- Check all my Docker containers and restart any that are unhealthy (it auto-heals on startup)
- Switch Kubernetes contexts, list pods, stream logs, scale deployments
- Run Ansible playbooks — with dry-run first, then confirm execution
- Check SSL certs across all my servers, run DNS lookups, port checks, traceroutes
- Read GitHub PRs, check workflow runs, get notifications
- Read my unread Slack messages from DigitalOcean workspace
- Ask "what changed last week?" and get a deployment timeline
The DevOps tools save me time, but the real value is everything in one place. I don't context-switch between terminals, dashboards, and Slack anymore. I just ask.
The Memory System
One thing I built early that I didn't expect to value so much: permanent memory.
MORPHEUS has a memory layer that survives chat clears and container restarts. Personal notes, task lists, trip plans, calendar events — all stored as files on my local machine, mounted into the Docker container. Every single OpenAI call injects these files directly into the system prompt.
# Tell MORPHEUS something once
"Remember: I prefer window seats on flights"
"Remember: my trading risk is 0.5% per trade"
"Remember: standup is at 11:30 AM, don't schedule trades before that"
# It writes to /morpheus-files/personal-notes/*.md
# And knows it forever — even after you clear the chatI've had it remember my Karachi trip plans, my gym schedule, my weekend reading list — and bring them up naturally in conversation without me asking. It feels like talking to someone who actually pays attention.
Kronos — AI Forecasting
MORPHEUS includes a separate AI forecasting service called Kronos. It takes live market data and generates a probabilistic view of where price is likely to go next. It doesn't make the final call — it's just another layer of context before I or the bot makes a decision.
What I Learned Building This
A few things surprised me:
AI tool calling is remarkably reliable. When you structure your tools well and give GPT-4o clear descriptions, it picks the right tool consistently. I've run hundreds of conversations and the routing is almost never wrong.
The AI confirmation layer matters. Adding an AI review step before any trade goes through was the single biggest improvement to bot performance. Not a new strategy — just an extra filter that asks "does this make sense right now?"
Personal context changes everything. When MORPHEUS knows I have a standup at 11:30, a Karachi trip coming up, and I'm running a 2% daily loss limit — it gives me genuinely different, better advice than a generic assistant would. Context at scale is what makes AI actually useful.
DevOps and trading are more similar than they look. Both are about managing complex systems, responding to signals, and staying disciplined under pressure. The mental models transfer directly.
What's Next
I'm actively working on the discipline layer — pre-trade gates that enforce risk limits before any order goes through, a formal daily loss limit with hard stops, and a weekly performance attribution report that breaks down P&L by strategy, session, and confluence combination.
The goal isn't to automate everything. It's to make every decision I take — in DevOps and in trading — more deliberate, more tracked, and better informed.
You don't need AI to replace you. You need AI to make sure you never make the same mistake twice.
See the Full Project Details
All features, architecture diagrams, screenshots, and bot details in the Lab.
View MORPHEUS in the Lab →