Fantasy Landscape

Smart AI. Smooth DeFi. The Future is MESSY.

Messy Virgo is an AI Fund OS that turns crypto noise into clear, followable rotations through research, allocation, and execution for crypto funds.

What we are building

Messy Virgo is an AI Fund OS: research tools feed decisioning, verification-phase funds prove the path, and public receipts keep the record falsifiable. Human-in-the-loop remains central. Fund managers can set the automation level.

Signal to sentence

The Glass Pipeline

From raw market signal to signed ledger mutation. Four transparent stages.

  1. 01 · Screening

    Research Engine

    Sensory surfaces aggregate into a shared daily substrate for screening and risk.

  2. 02 · Processing

    Cash · Base · High-beta

    Macro risk budgeting sets policy-bounded fund posture: how much cash, base, and high-beta exposure the book may carry this session.

    Cash Base High-beta
  3. 03 · Governance

    AI Council + human certification

    Multi-role agentic debate (Value, Macro, Trend, Exit, Risk, Chair) decides asset intent via Enter, Increase, Hold, Reduce, or Exit, then records an official resolution. A human certifies before anything mutates a book.

    Enter Increase Hold Reduce Exit
    See verification-phase funds ↓
  4. 04 · Operations

    Safety checks to signed ledger

    Planner, execution package, policy gates, and signing path turn certified intent into a ledger mutation. Under verification, a human stays on package review and the execution button.

    Gates → Ledger → Human sign

Core rule: Screening prepares evidence; Governance decides intent; Operations mutate positions.

Live proof · Full pipeline

Verification-phase funds

Where we are today: human-in-the-loop through research, bucket allocation, council, and gates & execution on Base micro test books. Not autonomous capital and not open for new LP deposits.

Base micro test funds

Verification Phase Guru Lotus custody

Live proving of the AI Fund OS on Base: screening, council decisioning, and execution under policy. Custody and share accounting for these books sit on Guru Lotus vault contracts. Messy runs the OS; provider-side contract risk is venue risk. These are small test books, not a product for new LP deposits.

Managed Guru.fund vehicles outside this test suite are being wound down. Do not treat historical Guru vault links as open Messy products.

Guru managed funds

Former Messy-linked Guru.fund vehicles (including $MESSYINFRA, $MESSYBNB, $MESSYAI, $MESSYDEFI and related managed books) are not open product. Outside the Base micro test suite, we are sunsetting them.

  • ·No deposit CTA from this site
  • ·Do not treat Guru vault pages as Messy “live funds”
  • ·Proving continues on labeled test books only
Latest Fund Update →

After verification proves the path

Toward manager-controlled automation

The four-stage pipeline above already runs under human review. What remains is productization and optional autonomy, not building Operations from scratch.

Coming

Allocation productization

Bucket allocation already runs in council prep. Next is hardening screening playbooks, thresholds, and risk rejection rules against real portfolios before broader automation options open.

Optional · manager-set

Configurable Autonomy

Verification defaults keep a human on package review. The AI Fund Agent is in development and not released. Unattended autonomy is optional and manager-set when the path is proven, not Messy’s default for every book. Do not treat verification-phase books as live AI-managed performance.

Updates & receipts

Where to read along: what shipped, what councils decided, what was rejected, and how the Fund OS hardens over time.

Institutional layer

Token and association trust sit apart from fund custody. Here: Swiss association execution, an audited token surface, renounced ownership, and public safes.

Live price

MESSY/WETH on Base via GeckoTerminal.

Verify & Buy $MESSY

Swap $MESSY on Aerodrome (Base). Use 4.95% slippage. Always verify the contract address before trading. Contract, safes, vesting, and association docs are public.

Join the community

Follow the latest findings on X and join Telegram for real-time discussions.