$ ./locusquant --boot

The operating system that runs
a hedge fund — agentically.

Multi-agent AI that researches, forecasts, and trades — governed by deterministic code gates, fully auditable, and proven against real outcomes before a dollar is at risk.

One continuous loop, from market feed to managed capital.

Every stage runs as coordinated agents under a central orchestrator. Outputs flow forward through deterministic gates, and open positions feed back into research as the fund is managed over time.

01
Live Market Feed
02
Thematic Segregation
03
Research & Stock Selection
04
Risk & Direction Gates
05
Execution & Capital Allocation
06
Continuous Fund Management
⟲ feedback loop: managed positions and realized outcomes flow back into Research & Stock Selection.

Three coordinated swarms.

Specialized agent groups, each owning a stage of the fund's operation and reporting through the orchestrator.

// swarm 01

Research Swarm

Enterprise-grade analyst agents working across historical data, market trends, and behavioral sentiment — grounded in established quantitative methodologies and technical indicators, not guesswork.

// swarm 02

Execution & Orchestration Swarm

Capital-deployment strategies for the selected equities, driven by the central orchestrator that sequences agents and routes every proposal through the risk and direction gates.

// swarm 03

Fund Management Engine

Portfolio-manager agents that keep ledger records and make state-based decisions — continuously rebalancing from the current as-is book toward the intended to-be allocation for profitability.

The model proposes. Deterministic code disposes.

What makes LocusQuant different from a chatbot pointed at a brokerage: the guardrails are code, not prompts — and every decision is on the record.

The edge isn't a magic signal — it's the layer that scores every forecast against reality, sizes down when it's uncertain, and proves its own skill on the record.

Deterministic code gates

Risk boundaries and house-view consistency are enforced by plain Python gates, not by prompt guidance a model can talk its way around. A non-compliant proposal is rejected before it can execute.

Append-only audit trail

Every agent output and gate decision streams to an append-only store, so each decision is traceable back to the inputs that produced it — no silent overrides.

Honest-by-default UI

Degraded desks, missing data, and gate interventions are surfaced, not hidden. The interface shows what actually happened, including when the system overrode the model.

A working prototype, today.

Not a mockup — a running multi-agent system you can watch deliberate, backtest, and audit.

// scope: today it runs a single ticker on paper capital — no live money — with every decision recorded and auditable.
// operational
Multi-agent deliberation — the Boardroom
Backtesting on a 252-trading-day basis — Strategy Lab
Portfolio & risk views
Full append-only audit log
// built on
LangGraph Groq-served Llama FinBERT SEC EDGAR FRED

Orchestrated multi-agent graphs over Groq-served Llama models, with FinBERT sentiment and grounded market & macro data from SEC EDGAR and FRED.

$ we automate research → prediction → deliberation → execution → monitoring — humans keep four irreducible seats: exception resolution, compliance judgment, capital allocation, and legal sign-off.

What ships next.

The concrete near-term releases from the pipeline — each one still paper-traded and routed through the same deterministic gates.

// next releases
v0.5 · in build Replatform — a real web platform: admin console, role-based access, kill switches, live Boardroom streaming, hash-chained audit store.
v0.6 Prediction Swarm — probabilistic forecasts scored against realized data; position size shrinks automatically as uncertainty rises.
v0.7 Execution & Breadth — from a single ticker to a screened universe, worked by a production-grade execution engine — paper first, always through the gates.
v0.8 Governance Layer — a compliance swarm audits every agent against the rulebook; humans steer agents through an instruction console.
v0.9 LocusOS — an immutable OS image runs every agent in its own micro-VM under an OS-level policy kernel no agent can route around.
v1.0 Fund-in-a-Box — the learning loop closes: trains on its own audit history, multi-book support. Install the image, get a hedge fund.

$ view the full pipeline → /whats-next.html

// the destination

A foundational model layered over the operating system — learning from the OS's own operational history to iteratively improve its strategies over time, so the fund gets sharper the longer it runs: the v1.0 “fund-in-a-box” destination.

Who's building it.

Divyanshu Bhardwaj
Data Science

Data science background across credit-bureau and fintech systems — the quantitative and modeling core of LocusQuant.

Ayush Vaishnav
Enterprise Integration

Enterprise integration background — wiring the operating system into robust, production-grade infrastructure.

$ investors and partners welcome — reach the team about investment opportunities in LocusQuant through the .