Cryptoholon

The Agent Stack

Six agents.
One trading floor.

Cryptoholon is not a single bot. It's a stack of six purpose-built agents, each with one job, feeding the next. Here's how a meme-coin trade actually moves through the system — from the first market signal to the closed session.

JUPITERRoutes.

Finds the best price across Solana liquidity for a trade you've already decided to make.

PHOTON · BULLXExecute.

Fast manual terminals. The trader is still the one deciding, pacing, and risk-managing.

CRYPTOHOLONDecides.

Six agents scan, filter, strategize, execute, monitor, and close — inside the parameters you set.

01 / The data flow

How a trade moves through the stack.

Every trade follows the same path. Each agent passes verified context to the next; nothing is implied.

  1. 00Live market data
  2. 01Market Scanner
  3. 02Opportunity Agent
  4. 03Strategy Engine
  5. 04Execution Engine
  6. 05Position Monitor
  7. 06Session Controller
  8. ∑Session history

02 / The Agent Stack

Six agents, each doing one job.

Separation of concerns is the architecture. Every agent has clear inputs and clear outputs. If a trade looks wrong, you can trace it to the exact agent that made the call.

01

Market Scanner

Continuously watches supported chains for new activity, liquidity, and volume shifts.

The Market Scanner is the sensor layer. It subscribes to live onchain feeds across Solana, Ethereum, Base, BNB Chain, and Arbitrum, watching for new pairs, liquidity movements, and volume spikes. The Scanner does not decide anything — it reports. Everything downstream depends on the Scanner never missing a condition that could matter.

Inputs
Live chain feedsPriceLiquidityPair ageVolume
Outputs
Normalized market snapshots
02

Opportunity Agent

Flags conditions worth evaluating before they trend — not after.

The Opportunity Agent reads the Scanner's snapshots and asks a single question: is this condition worth evaluating? It filters out pairs that are too thin, too new, too old, or too quiet to be actionable. Pairs that survive the filter become candidate opportunities and get passed to the Strategy Engine. The Opportunity Agent does not place trades.

Inputs
Market snapshotsLiquidity floorVolume floorPair-age limits
Outputs
Candidate opportunities
03

Strategy Engine

Tests opportunities against the active strategy and your session rules.

The Strategy Engine is where opportunities become decisions. It runs each candidate opportunity against the strategy you selected for the session — momentum, mean-reversion, breakout, or whichever strategies are enabled on your account — and checks whether a trade is justified under your session parameters. If the strategy says yes and your rules allow it, the Strategy Engine instructs the Execution Engine to open a position.

Inputs
CandidatesStrategy rulesSession parameters
Outputs
Trade instructions
04

Execution Engine

Handles order routing, slippage control, and onchain settlement.

The Execution Engine takes a trade instruction and makes it real. It routes orders through the best available liquidity, controls slippage inside configured bounds, and settles onchain. It also reports back exactly what filled, at what price, and at what cost — so every subsequent decision uses actual data, not intended data.

Inputs
Trade instructionsRouting policySlippage ceiling
Outputs
FillsOnchain receipts
05

Position Monitor

Tracks open positions, PnL, and exit triggers in real time.

Once a position is open, the Position Monitor owns it. It tracks the position's PnL against the market in real time, watches for the exit triggers the Strategy Engine set, and reports position state back to the Session Controller. When an exit condition fires — target, stop, or session close — the Monitor hands the exit instruction to the Execution Engine.

Inputs
FillsLive market dataExit triggers
Outputs
Exit instructionsPnL state
06

Session Controller

Enforces the risk parameters and limits you selected. Nothing exceeds your plan.

The Session Controller is the discipline layer. It enforces your capital allocation, maximum concurrent positions, per-trade size, risk ceiling, and session duration. The other agents propose; the Controller decides whether a proposal is allowed. When the session's capital, time, or risk limits are reached, the Controller closes the session cleanly and preserves the full history.

Inputs
Session parametersLive PnLAgent proposals
Outputs
Allow/deny decisionsSession close

The agents propose.
The Controller decides.

— How Cryptoholon enforces discipline

03 / Session controls

You set the shape. The agents work inside it.

A trading session is the container. Before it starts, you define every parameter that bounds the agents' behavior. The Session Controller enforces them in real time.

NEW SESSIONTrading session setup
Ready to deploy
500 USDT
How much of your trading balance commits to this session.
SolanaEthereumBaseBNB ChainArbitrum
Which supported chains the agents may scan and trade on.
Momentum · high-volatility
Which strategy the Strategy Engine evaluates opportunities against.
5% per trade · 3 concurrent
Max exposure per trade and max concurrent positions.
4 hours
A clear start and end — no runaway sessions.
25K liquidity · 50 volume/h
Floors the Opportunity Agent applies before passing a candidate on.
Example setup — actual values are yours to choose each session. Capped risk means chosen parameters. It does not mean a guaranteed loss limit or a guaranteed return.

More questions?

The FAQ covers the rest.

Fees, custody, chain support, minimums, strategies, how we compare to Jupiter and Photon — the home FAQ has 12 answers.

Open the FAQ

Ready?

Set the parameters.
Let the agents trade.