Skip to main content
Grid mode places limit orders at configured levels around the current mid price. The agent manages the grid automatically and rebuilds it when the market moves beyond the configured threshold.
Bids are placed below the mid price using negative offsets, while offers are placed above it using positive offsets.

Before you start

Complete agent onboarding first. It covers installing cloud-agent, setting up your agent identity, and creating the initial .env and agent.toml files.

Configure grid mode

1

Set the agent configuration

Open the agent.toml created during onboarding and check the main agent settings:
Keep rfq_v2_only = false when using grid orders.
2

Add a market

Add the market where you want the agent to place grid orders:
price_change_threshold_percent defines how far the mid price must move before the current grid is canceled and rebuilt at updated levels.
3

Add bid levels

Configure bids below the current mid price using negative delta_percent values:
Each quantity defines the order size for that level.
4

Add offer levels

Configure offers above the current mid price using positive delta_percent values:
Once the market and levels are configured, the grid is ready to run.

Run the grid

Choose how you want the agent to operate:
Run only the grid-order workflow:
This mode places and manages grid orders without processing settlement.
Grid orders are not placed when rfq_v2_only = true. The --orders-only option also conflicts with this setting.

Test your configuration

Use --dry-run to prepare and verify transactions without signing or executing them:
Once the configuration behaves as expected, start the agent without --dry-run.

Useful options

Each grid level is defined by:
  • delta_percent — percentage offset from the current mid price
  • quantity — order size at that level
When the mid price moves by more than price_change_threshold_percent, the agent cancels the existing grid orders and places them again at updated price levels.
For more advanced configuration, CLI options, and implementation details, see the Silvana Book Agent README.