See how an agent behaves before it can act.
Choose a strategy, set the parameters, and watch the agent respond to a simulated market. Export the configuration when you are ready to continue in the developer workflow.
Simulation only. No funds. No live orders. No production activity.
A controlled place to understand the strategy.
The Playground is not a trading product and it does not connect a strategy directly to production. It is an interactive explanation of how an agent responds to inputs, constraints, and events.
Configure the mandate
Choose the strategy and define the parameters that shape its behavior.
Watch each action
Follow placements, fills, adjustments, and simulated settlement events as they happen.
Understand the reason
See which signal, rule, or limit caused the agent to take each action.
Export the result
Leave with a configuration that can become the starting point for the SDK workflow.
Set the mandate before the agent starts.
The simulator begins with a small set of meaningful controls. Each parameter is visible before the simulation runs.
Strategy
Choose the operating pattern the agent will follow. The current simulator supports grid trading and market making.
Simulated market
Use CC / USDC as the mock pair for the session.
Spread
Set the distance between the agent's buy and sell prices. The default is 0.4 percent.
Price levels
Choose how many order levels the agent maintains on each side of the market. The default is 6.
Order size
Set the simulated quantity placed at each active level. The default is 250.
Two patterns to inspect first.
Grid Trading
The agent places buy and sell orders at defined price levels, then updates the grid as the simulated market moves or a fill occurs.
Market Making
The agent maintains bid and ask quotes around a simulated reference price while keeping the configured spread and size in view.
Follow the market, the decision, and the action.
The workspace makes the agent's behavior legible at a glance. It should feel like a controlled instrument rather than a decorative trading screen.
Market panel
Shows the mock orderbook, reference price, bid and ask levels, and simulated market movement.
Agent status
Shows whether the agent is preparing, placing, monitoring, adjusting, paused, or complete.
Active orders
Shows the simulated orders currently maintained by the strategy.
Position summary
Shows the simulated inventory created by fills during the session.
Decision log
Explains why the agent placed, changed, or removed an order.
Event timeline
Records market changes, order events, fills, adjustments, and simulated settlement outcomes in sequence.
Every event should have a visible cause and result.
Read the market
The agent receives the current mock price, spread, depth, and active order state.
Evaluate the strategy
It compares those inputs with the selected spread, levels, order size, and current inventory.
Place the orders
It creates the simulated orders required by the selected strategy.
React to events
It responds when the market moves, a fill occurs, or an active level is no longer valid.
Rebuild the state
It replaces or recenters orders so the strategy returns to its intended structure.
Record the outcome
The simulator updates the position, decision log, event timeline, and exportable configuration.
Show the reason, not only the result.
Every meaningful action should open into the same five part explanation.
Signal
What changed in the market or current state?
Decision
Which strategy rule became relevant?
Action
What did the agent place, cancel, replace, or update?
Constraint
Which spread, size, level, inventory, or timing rule shaped the action?
Outcome
What changed in the active orders, position, or simulated workflow?
Example event
When the simulated reference price moves beyond the current grid center, the agent removes stale levels and places a new set around the updated reference price.
Grid recentering
Use short, specific event messages.
Market events
Reference price moved, spread widened, spread narrowed, or depth changed.
Agent events
Strategy initialized, orders placed, grid recentered, quotes refreshed, paused, or resumed.
Fill and settlement events
Buy order filled, sell order filled, position updated, or DvP settlement completed in simulation.
Completion event
Simulation complete. Review the behavior or export the configuration.
The simulation ends with something useful.
The export translates visible controls into a clear developer handoff. Nothing deploys automatically from the Playground.
Configuration file
Download the generated agent.toml as the starting point for the SDK workflow.
CLI command
Copy the command required to run the configuration in the supported developer environment.
Strategy summary
Review the selected strategy, market, spread, level count, order size, and simulation result.
Simulation report
Download or copy the event sequence when teams need a record of the test session.
A structural starting point.
The final configuration must match the current SDK schema before publication. Synchronize field names, accepted values, commands, and file structure with the maintained SDK documentation before launch.
[market]
pair = "CC/USDC"
mode = "simulation"
[strategy.grid]
spread_percent = 0.4
levels = 6
order_size = 250
[signing]
mode = "local"Move forward without treating simulation as production.
Each environment has a different purpose and a different level of consequence.
Playground
Use mock market data to understand strategy behavior and generate an initial configuration.
Developer Sandbox
Run the configuration against supported test services and test assets before production deployment.
Production Deployment
Review the configuration, connect required services, choose infrastructure, and launch through the SDK workflow.
Describe the strategy. Start from a reviewable draft.
Natural language builder
A future workflow can turn a plain language strategy into an editable configuration or SDK starting point. The generated result must always remain visible, reviewable, and subject to explicit approval.
A simulation explains behavior. It does not promise performance.
The Playground can help you
The Playground does not prove
The Playground is an educational and development simulation. It does not place live orders, use real funds, or provide a forecast of future performance.
Keep every state understandable.
Ready
Configure the strategy and start the simulation when you are ready.
Preparing
Building the mock market and initial order state.
Running
The agent is monitoring the simulated market and applying the selected strategy.
Paused
The simulation is paused. Resume when you are ready to continue.
Complete
Review the timeline or export the configuration.
Reset confirmation
The current event history and simulated position will be cleared.
Error
Reset the session and try again if the simulation stops before completion.
Reduced motion
Use discrete state changes instead of continuous animated movement when reduced motion is enabled.
Test the behavior. Then decide what deserves to run.
Use the Playground to understand the strategy, then continue with the SDK when you are ready to build and deploy.