Agentic Finance · OpenClaw Teams

An AI-native hedge fund that actually does the research.

Meet Victor — OpenClaw's AI portfolio manager. He combs 10-Ks, earnings calls, and SEC filings, runs real technical & fundamental analysis, and proposes disciplined, risk-limited trades — all inside Telegram or your browser console.

For educational & research purposes only. Nothing on this page constitutes financial advice, a solicitation, or a recommendation to buy or sell any security. Always consult a licensed financial professional before trading.

Start in Telegram Open Web Console

Real SEC EDGAR 10-K citations — not hallucinated summaries

Risk engine enforces ≤5% per name, 15% cash floor, human gate

Paper-account by default — live trading only with your broker creds

See Victor in action

Live-simulated conversation — the numbers and citations below are real, validated agent outputs. Switch between Telegram and Slack UI skins.

Demo loops automatically. Figures are real validated outputs — not fabricated. Not financial advice.

How Victor works

Every call Victor makes is grounded in four real data layers — no LLM-invented numbers.

Live price & technicals

52-week range & real drawdown
SMA50 / SMA200 crossovers
RSI14 overbought / oversold
Annualised volatility — σ-normalises every move
Source: Yahoo Finance real-time feed

Fundamentals

SEC EDGAR 10-K annual filings
Citations by accession number (e.g. 0000909832-25-000101)
Revenue, net income, margin trends
Balance sheet red-flag scan
Source: SEC EDGAR XBRL data

Risk engine

≤5% NAV per name (hard cap)
15% cash floor — always maintained
25% per-sector concentration cap
Daily kill-switch on drawdown breach
Human approval gate before any live order

Versions & how it evolves

v1 ships today as a disciplined PM. v2 turns it into a real, coordinating agent team.

v1● Live

Solo PM & desks

A disciplined single portfolio manager — optionally backed by a research desk and a paper-trading desk. Real data, a hard risk engine, and simulated execution on a paper account. Every behaviour was hardened by a live evaluation loop, not by assumption.

v2● In design

Real agent team

Specialists that each have their own tools, skills, and memory, coordinating through a defined handoff matrix with real ownership and veto rights — provable live delegation, where the team beats a solo PM on multi-domain calls.

Battle-tested by live evaluation

Each behaviour below was caught by a real eval against the live agent and fixed at the root — the kind of failure a unit test never surfaces.

Never fabricates a price or a SEC accession number — says “couldn’t fetch” instead

Every recommendation carries an explicit “not financial advice” disclaimer

Flags the $25k pattern-day-trader rule for small day-trading accounts

Adapts strategy to the goal: long-term low-turnover vs high-risk day-trade

Hand it a portfolio and it does the work — no cold-start “who am I” intake

Pick a preset & deploy

Choose how much team you want — from a solo PM to a full paper-trading desk. You confirm and launch it in the console; paper account by default.

Roster — Paper-Trading Desk

PMPortfolio Manager· aka Victor
EREquity Researcher· aka Priya
CRCrypto / DeFi Researcher· aka Kai
RARisk Analyst· aka Morgan
CCompliance· aka Lex
EExecution Broker· aka Apex

What it does

Everything in Research Desk
Execution agent runs simulated orders (SimulatedBroker)
Paper account — no real money by default
NAV / Sharpe / drawdown vs SPY tracking

Trading mode

Deploy Paper-Trading Desk

Configure & confirm in the console. Paper account by default · not financial advice.

Deploy your own AI portfolio team

Victor runs on OpenClaw's managed agentic infrastructure — private, isolated, paper-account by default. Start with one /start in Telegram.

Start in Telegram Open Web Console See All Agent Teams

For educational purposes only. Not financial advice. Past simulated performance does not guarantee future results. OpenClaw is not a registered investment adviser, broker, or dealer.