Agent types

Three ways in. Pick your depth.

The same agents, the same Canton settlement. What changes is how much of the machinery you want to hold yourself — from one button to your own validator.

Open the agent app Technical fundamentals

Agent apppress deploy

Agent templatesfork a crate

Silvana SDKcrates and a CLI

Agentic APIthe ledger, over gRPC

Your validatoryour own premises

For individual traders
One button. Nothing to install, and no code.
For developers
Fork a crate. Open source, down to the crate.
For trading firms
All five. Your own validator, on your own premises.

Agent apppress deploy

Agent templatesfork a crate

Silvana SDKcrates and a CLI

Agentic APIthe ledger, over gRPC

Your validatoryour own premises

For individual traders
One button. Nothing to install, and no code.
For developers
Fork a crate. Open source, down to the crate.
For trading firms
All five. Your own validator, on your own premises.

By who you are

Start where you actually stand.

Three audiences, and a different thing on offer to each. The right panel is the one that describes what you did yesterday.

For individual traders

Your strategy stops when you do. It does not have to.

The strategy you already run, executing through the hours you are not at the desk — with a model watching it work and recalibrating as conditions move, rather than a rule repeating itself into a market that has changed.

Agent App
Agent Templates

Automate what you already do

Your entries, your sizing, your limits. The agent runs them without you in the chair, and it does not get bored at four in the morning.

A model keeps them honest

The strategy still executes. What changes is that something is watching it execute, with the authority to recalibrate — or to stand it down when the conditions it was built for are gone.

Take the hours back

The gain is not that a machine trades faster than you. It is that watching a chart stops being the price of holding a position.

Open the agent app

Hours a day the book is covered: 8 at the desk24 with an agent
Setup before the first run: Minutes — nothing to install, no code

For trading firms

Staff the desk that never clocks off.

Whole functions — execution, risk, treasury, compliance — running as agents on Silvana's stack, so the operational work between a decision and a settled trade stops being headcount.

Enterprise Agent Solutions

Departments, not scripts

The six agent families map onto desks a firm already has. Execution works the order, risk holds the limits, compliance proves the run afterwards.

Built on what already runs

Silvana's agents, SDK and catalogue templates are the floor. The firm's own logic sits on top of them instead of underneath a year of plumbing.

Your validator, your premises

Transactions through your own Canton validator, custom changes on top of the agent crates, and an execution environment you hold end to end.

Talk to Silvana

The desks, staffed

Execution
Strategy
Portfolio
Risk and control
Oracle and data
Compliance and audit

On the board

CBTC-USDCxBIDExecutionRisk and control

cETH-USDCxASKStrategyOracle and data

CC-USDCxBIDPortfolioCompliance and audit

Agent families you can staff: All six — execution through compliance
Where execution sits: Your own validator, on your own premises

For developers

Read the agent, then make it yours.

The SDK and the catalogue templates are working code rather than samples. Take the nearest one and change what you disagree with — the crates are open source, and the fastest way to understand the SDK is to read an agent built on it.

Silvana SDK
Agent Templates
Documentation
Utilities

Start from something that already runs

Every catalogue entry ships a crate, a doc.md and a demo you can open. A template is a working starting point, not a sketch to fill in.

Change anything

Open source down to the crate. Fork it, rewrite the strategy, swap the model, keep the parts that were never the interesting bit.

More than the SDK

The typed gRPC surface, the CLI, reflection, the configuration and operation references, the security model — callable from any language with a gRPC client.

Technical fundamentals

agent-spot-grid-static/src/main.rs

use anyhow::{Context, Result};

use tokio::sync::Mutex as TokioMutex;

use agent_logic::client::OrderbookClient;

use agent_logic::config::BaseConfig;

use agent_logic::runner::{run_agent, AgentOptions};

use orderbook_proto::orderbook::OrderType;

use cloud_agent::CloudSettlementBackend;

let config = BaseConfig::load(&cli.config)

.with_context(|| {

format!("Failed to load {:?}", cli.config)

})?;

let enabled_markets: Vec<&str> = config

.enabled_markets()

.iter()

.map(|m| m.market_id.as_str())

.collect();

if enabled_markets.is_empty() {

anyhow::bail!(

"agent.toml has no enabled [[markets]] blocks"

);

}

info!(

"Enabled markets: {}",

enabled_markets.join(", ")

);

// Enabled markets: CBTC-USDCx, cETH-USDCx

let mut client =

OrderbookClient::new(&config).await?;

for m in &client.get_markets().await? {

let parts: Vec<&str> =

m.market_id.split('-').collect();

liquidity

.register_alias(parts[0], &m.base_instrument)

.await;

}

use anyhow::{Context, Result};

use tokio::sync::Mutex as TokioMutex;

use agent_logic::client::OrderbookClient;

use agent_logic::config::BaseConfig;

use agent_logic::runner::{run_agent, AgentOptions};

use orderbook_proto::orderbook::OrderType;

use cloud_agent::CloudSettlementBackend;

let config = BaseConfig::load(&cli.config)

.with_context(|| {

format!("Failed to load {:?}", cli.config)

})?;

let enabled_markets: Vec<&str> = config

.enabled_markets()

.iter()

.map(|m| m.market_id.as_str())

.collect();

if enabled_markets.is_empty() {

anyhow::bail!(

"agent.toml has no enabled [[markets]] blocks"

);

}

info!(

"Enabled markets: {}",

enabled_markets.join(", ")

);

// Enabled markets: CBTC-USDCx, cETH-USDCx

let mut client =

OrderbookClient::new(&config).await?;

for m in &client.get_markets().await? {

let parts: Vec<&str> =

m.market_id.split('-').collect();

liquidity

.register_alias(parts[0], &m.base_instrument)

.await;

}

$ silvana agent run --config agent.toml

Templates to fork: 68 crates, each with a demo and a doc
How far you can change them: Open source, down to the crate

By job

Six shelves, sixty-eight agents.

Whatever you were about to do by hand at two in the morning, something in here already does it.

Execution

Work an order without moving the market against yourself.

TWAP Execution
Iceberg Execution
Algo Order
Liquidity Seeking

Strategy

Take a position on a rule you can state and test.

Spot Grid
Trend Following
Pairs Trading
Mean Reversion
Arbitrage Agent

Portfolio

Hold a shape across instruments instead of one trade at a time.

Portfolio Rebalancing
Target Allocation
Smart Allocation
DCA Portfolio

Risk and control

The agents that stop the other agents.

Killswitch
Circuit Breaker
Risk Management
Pre-Trade Check
Concentration Risk

Oracle and data

Bring outside truth on-chain, and publish what happened.

Oracle
Fair Value
Orderbook Streaming
Volatility Screening

Compliance and audit

Prove the run, not just the result.

Audit Attestation
Audit Replay
Compliance Screening
Selective Disclosure

The shelves are ours; the agents are not marketing. Every name above is a live entry in the catalogue, with a template, a demo and a repository behind it.

By what runs it

Four kinds of thing can hold a job.

Independent of the shelf above: a portfolio job can be plain automation or a model in an enclave. This is what is doing the work, not what the work is.

Automation agents

Sequence transactions and app operations — request handling, account abstraction, flow automation, and scheduled, pre-paid, delayed or event-triggered transactions.

Deterministic. Given the same inputs it does the same thing, every time.

Reached fromStandardPro

Oracle and witness agents

Witness and fetch on- and off-chain business data, and publish it provably so the rest of the workflow can rely on it.

Brings in facts from outside the ledger, with a proof attached.

Reached fromProEnterprise

AI agents

Run inside secure enclaves to produce insights and encrypted interpretations of data the raw numbers do not state outright.

Judgement. This is the archetype that can decide the rule no longer applies.

Reached fromStandardProEnterprise

Human agents

Professionals and individuals providing manual checks, audits or expert opinion at the points a workflow calls for one.

A person, addressed through the same interface as the code.

Reached fromEnterprise

Try before anything

Open one and watch it work.

These four are running demos. No wallet, no signup, no install — just the agent doing its job in a browser tab.

Arbitrage Agent

Earn the gap between venues while you sleep.

Watches the same pair on CEX and Canton DEX venues at once and takes the spread when it opens wider than your floor.

Open the demo Source

Portfolio Rebalancing

Hold your target weights without watching them.

Values the book each cycle and trades the drift back out, so the split you chose is the split you keep.

Open the demo Source

Spot Grid

Let a sideways market pay you for the chop.

Stacks bids and offers at fixed steps around the mid and lets price walk through them. No chasing, no RFQ.

Open the demo Source

Trend Following

Be in before the move is obvious.

EMA crossover momentum — enters when fast crosses slow, steps aside when it crosses back. Never against its own position.

Open the demo Source

Now the how.

The SDK, the gRPC surface, the crates and the docs — everything a developer needs to take one of these apart and build their own.

Technical fundamentals Browse live agents