Monstra loading art

Strategy Platform

Loading Monstra

Preparing the next route...

MonstraTM
  • Home
  • About
  • Monstra Bots
  • Vectors
  • Titans
  • Armor
  • Heartbeat
MonstraTM

Monstra is a simulation and educational platform. Bot signals are algorithmic outputs, not personalized advice. Users make their own decisions, and past simulated performance does not guarantee future results.

Product

  • Bots

Explore

  • About
  • Blog
  • Support

Legal

  • Terms of Service
  • Privacy Policy
  • Risk Disclosure
  • Brokerage Disclosure

© 2026 Monstra. All rights reserved.

Built for research, simulation, and strategy exploration on Monstra's dark-first platform.

Algorithm

Oracle

Sends live market context to an LLM each cycle and proposes an allocation, but never trusts it directly -- a code-enforced policy gate re-derives every constraint before anything reaches a trade.

Oracle Bots

Bellitus

Bellitus artwork

Creator

@official

Type

Oracle

Bellator

Bellator artwork

Creator

@official

Type

Oracle

Cyclus

Cyclus artwork

Creator

@official

Type

Oracle

Imperium

Imperium artwork

Creator

@official

Type

Oracle

Metalli

Metalli artwork

Creator

@official

Type

Oracle

Medicus

Medicus artwork

Creator

@official

Type

Oracle

Vectura

Vectura artwork

Creator

@official

Type

Oracle

Vis

Vis artwork

Creator

@official

Type

Oracle

Cycli

Cycli artwork

Creator

@official

Type

Oracle

Regina

Regina artwork

Creator

@official

Type

Oracle

Cuprum

Cuprum artwork

Creator

@official

Type

Oracle

Vitalis

Vitalis artwork

Creator

@official

Type

Oracle

Rapax

Rapax artwork

Creator

@official

Type

Oracle

Dux

Dux artwork

Creator

@official

Type

Oracle

Thesauro

Thesauro artwork

Creator

@official

Type

Oracle

Annona

Annona artwork

Creator

@official

Type

Oracle

How Oracle Works

Oracle

Oracle in 4 steps

Send an LLM the bot's candidate universe and current holdings, let it propose an allocation with a reason for each pick, then run that proposal through a code-only policy gate that never trusts the model directly — only a proposal that survives every rule reaches a trade.

1Context
MSFTmomentum0.82
VLOtrend0.77
JPMquality0.64

Candidate tickers, tag scores, current holdings.

2Propose

Proposal

MSFT 28% · VLO 22% · JPM 18%

“Top-ranked for momentum and trend, balanced with a quality name for stability.”

The model proposes tickers, weights, and a rationale.

3Policy gate
✓In candidate universe
✓Within min/max holdings
✓Under per-name weight cap

Code re-checks every rule before anything trades.

4Trade or hold

Gate result

Accepted

Rejected or malformed proposals hold — never a fabricated fallback trade.

Rejected or unclear proposals hold the current position.

Example only. Tickers, returns, scores, and weights shown here are illustrative, not live signals or trading advice.

Oracle sends an LLM its candidate universe, tag scores, and current holdings each cycle and asks it to propose an allocation with a rationale — then re-validates that proposal in code before anything ever trades. The following slides walk through a real, gate-accepted example. Oracle is currently available as an official bot only, not yet open to custom user-built configurations.

  • An LLM proposes tickers, weights, and a short rationale from a closed candidate list -- it cannot name a ticker outside that list
  • The proposal is never trusted directly: a deterministic policy gate re-derives every constraint from scratch
1 / 4