Monthly returns
Strategy report · 2025
- Jan+0.2%
- Feb+2.0%
- Mar+2.8%
- Apr+2.8%
- May−0.8%
- Jun−3.9%
- Jul−1.6%
- Aug+2.8%
- Sep+5.2%
- Oct−1.9%
- Nov+5.1%
- Dec+4.0%
Year 2025+16.7%
An entire trading desk, powered by AI — scalable cloud compute and optimisation algorithms searching, validating and deploying strategies at a scale no human desk could match.
Illustration of a VC Trader session. The trader asks Agent Alpha to review FX and metals for the session, choosesClaude Opus 5.5 and sends the request. Agent Alpha reads the desk, the IC Markets MT5 account, the event calendar, live quotes and its research memory, briefs three helpers, reruns a gold strategy backtest in the cloud and charts XAUUSD with structure, a buy area, invalidation and a target, followed by a written analysis. The trader then asks it to place the trade; Agent Alpha refreshes the account and market, sizes and validates the order, and presents a buy-limit proposal awaiting the trader's approval. All figures are illustrative.
A proprietary agentic harness purpose-built for trading operations. Coordinate specialist agents, define their workflows and connect research, market analysis and live execution in one continuous workspace.
Agents prepare and execute trades with swipe-to-confirm approval for agent-proposed orders.

Gold specialist
ActiveXAUUSD · Claude Sonnet 5.5
Wakes every 5 min
AlertXAUUSD crosses 4,150
Instructions
Run on wakeAgent Alpha · trade proposal
XAUUSD Buy limitIC Markets MT5 · demo
0.05 lots · $85 at risk · 0.34%
Awaiting approval
Situational Awareness
VC Trader's proprietary Situational Awareness brings watched markets, multi-timeframe conditions, economic events, account exposure and recent decisions into a unified read. Your agents work with the state of the desk and the context around it.
Situational awareness
LiveBook · 2 accounts
Same bet: both positions are short USD
computed from live positions and their currency legs
Calendar · USD
FOMC rate decision
Federal Reserve · in 2 days
EURUSD and USDCHF moved inversely on 86% of days
Agent Alpha · observation
Your EURUSD long and USDCHF short are the same bet: both are short USD, about 30.9k combined. The FOMC decision is the next high-impact USD event.
Proposed Review combined USD exposure before the release.
Awaiting your review · nothing changed
A semantic intelligence graph connects companies, markets, currencies, strategies, research, decisions and outcomes. Agents traverse those relationships to uncover hidden connections, retrieve relevant findings and bring accumulated knowledge into your trading desk.
Short USD 10.9k via EURUSD
486 of 982 trades in the London session
Hold above 4,150 confirms trend
Approved 5 Oct · closed at target
Desk watchlist · 7 instruments
Meta · evidence connected
METASource documentResearchApple · evidence connected
AAPLSource documentResearchNVIDIA · evidence connected
NVDASource documentResearchMicrosoft · evidence connected
MSFTSource documentResearchAmazon · evidence connected
AMZNSource documentResearchAlphabet · evidence connected
GOOGLSource documentResearchShort USD 10.9k via EURUSD
486 of 982 trades in the London session
Hold above 4,150 confirms trend
Approved 5 Oct · closed at target
Desk watchlist · 7 instruments
Meta · evidence connected
METASource documentResearchApple · evidence connected
AAPLSource documentResearchNVIDIA · evidence connected
NVDASource documentResearchMicrosoft · evidence connected
MSFTSource documentResearchAmazon · evidence connected
AMZNSource documentResearchAlphabet · evidence connected
GOOGLSource documentResearchResearch note (3 Oct) → about → XAUUSD, and → supersedes → the 18 Sep finding
Explore performance, understand risk and examine strategy robustness with account analytics, backtest reports and walk-forward diagnostics—all in one workspace.
CMA-ES search
RunningMTF_Breakout v8 candidate · XAUUSD H1
AWS research compute · fill model: tick-accurate
atr_mult · breakout_len · session_filter
session_edge.ipynb
Python 3 · idleimport pandas as pd
oos = pd.read_parquet("oos_trades.parquet")
by_session = (
oos.groupby("session")["r_multiple"]
.agg(trades="count", mean_r="mean",
win_rate=lambda r: (r > 0).mean())
.round(2)
)
by_sessionby_session = oos.groupby("session")
by_session["r_multiple"].agg(
trades="count", mean_r="mean"
).round(2)| session | trades | mean_r | win_rate |
|---|---|---|---|
| Asia | 212 | -0.04 | 0.46 |
| London | 486 | 0.21 | 0.58 |
| New York | 284 | 0.09 | 0.53 |
Rolling walk-forward
Train (in-sample) Test (out-of-sample)
Quant Research
Bring strategy development, custom analysis and large-scale experimentation into one environment. Develop in Python, test with tick-level precision, and put scalable cloud compute behind parameter optimisation and deeper analysis of your results.
Cloud Deployment
A scalable hosting environment for your trading operations. Run Python strategies on managed Linux infrastructure, maintain secure connections to your brokers and host Windows MT5 terminals where needed—bringing strategy deployment, account connectivity and operational oversight into one workspace.
Customer-approved Python strategies trade automatically. Orders an agent proposes still wait for your confirmation.
MTF_Breakout v7
Deployed 3 Oct · Python strategy

Turn your risk rules into operational controls. Set account limits, monitor exposure across deployments and check orders before execution. Combine circuit breakers and a kill switch with detailed analytics and a trading journal to oversee activity and intervene when needed.

Kill switchHalts trading and flattens all positions
Account limits
Risk gate · pre-trade check
XAUUSD buy limit · 0.05 lots · stop 4,121.0
Within your limits · awaiting your approval
Trading journal
Models & BYOK
Our proprietary Harness turns model intelligence into a connected trading desk, equipping agents with research and execution tools, configurable workflows and live operational context. Through the knowledge graph, they retrieve prior findings, follow relationships and build on the accumulated knowledge of your operation.
Model · Agent Alpha
BYOKAnthropic
DeepSeek
Same harness: tools, workflows and workspace context
FAQs
They can research markets, coordinate specialists, develop and test strategies, monitor your workspace and prepare trades for live execution. Agent-proposed trades require swipe-to-confirm approval.
You can work with agents to author Python strategies, or write and investigate them yourself in Jupyter. The research environment supports both approaches.
It connects research to instruments, strategies, decisions and outcomes through semantic relationships. Agents can retrieve relevant findings and carry context between investigations.
Tick-accurate backtests, CMA-ES parameter searches and rolling walk-forward evaluation, alongside Python analysis in Jupyter notebooks.
Yes. Customer-approved Python strategies can execute automatically on connected MT5 accounts through cloud hosting. Agent-proposed orders require your confirmation.
Yes. VC Trader supports model choice and BYOK within the same agentic harness and workspace.
Explore the platform through paid early access.
Apply for early access