Unlocking Autonomy on Canton: The Agent Automation Kit

Canton and Daml are the gold standards for defining rights, obligations, and privacy. But smart contracts are inherently passive: they do not wake up on their own to fetch oracle data, settle expired markets, or execute recurring payments.

To operate at scale, they need a secure way to act.

Today, Autonomous Finance is announcing the Canton Agent Automation Kit: a practical, production-ready framework to turn static applications into self-executing financial engines.

The Problem: The Automation Gap

Real financial applications are dynamic. They require action after time passes, conditions are met, or market data arrives. Generic cron jobs are not enough for Canton. True autonomous finance requires party-aware authorization, ledger-defined state, explicit grants, and execution receipts.

Until now, developers had to build custom middleware to trigger these events, often resulting in fragile, siloed, and opaque infrastructure. The Canton Agent Automation Kit standardizes this. It provides the offchain architecture needed to safely observe the ledger, claim tasks, and submit authorized commands without manual intervention.

The Next Evolution of Canton Automation

The Canton ecosystem has long explored ways to automate ledger state. Developers are already familiar with native approaches like Daml Triggers, utility tools like Daml Script, and custom-built Ledger API clients.

While these tools paved the way, scaling autonomous applications requires a robust, production-grade standard. Building highly available, oracle-driven agents from scratch takes focus away from core application logic and forces teams to engineer custom middleware.

The Canton Agent Automation Kit bridges this gap. It standardizes the Ledger API architecture, delivering reusable, production-ready offchain runners so you can deploy secure automation without building the underlying infrastructure from the ground up.

What We Are Shipping

This is not just theory. It is a comprehensive toolkit designed to keep the ledger as the ultimate source of truth while your offchain agents act as constrained, highly efficient workers.

• Ledger Primitives: Reusable smart contract patterns that safely govern agent authority, manage task lifecycles, and generate verifiable execution receipts.
• Offchain Runners: Standardized reference architecture for observing state and securely submitting authorized commands back to the ledger.
• Production-Ready Ops: Deployment blueprints, operational safeguards, and best practices for scaling safely in enterprise environments.
• Field-Tested Guidance: Comprehensive documentation and instructions from the field on how to design, ship, and maintain autonomous agent flows securely.

Hardened in Production: Nous and Upline

We are not building this in a vacuum. We are stress-testing the kit against real-world, high-velocity applications:

• Nous: Powering AI-driven prediction agents that turn complex offchain data into instant ledger action.
• Upline: Autonomously resolving 5-minute up/down markets. Agents pull data from Pyth and Chainlink, submit auditable evidence, and trigger settlement entirely hands-free.

If it works for high-frequency, time-sensitive market resolutions, it will handle your enterprise workflow.

Unlocking New Primitives for DeFi and Beyond

Our vision extends far beyond prediction markets. The Canton Agent Automation Kit is designed to establish autonomous agents as a foundational primitive. By standardizing how agents securely interact with the ledger, we are unlocking a massive design space for developers.

These tools do not stop with our initial use cases. They are built to be adapted by teams across decentralized finance and traditional enterprise architecture. Whether you are building programmable payment rails, dynamic liquidity management, cross-chain custody workflows, algorithmic risk assessments, or automated governance execution, we are providing the composable building blocks so the broader ecosystem can scale autonomous workflows into new frontiers.

The Reference Agents

Out of the box, developers get pre-configured agents for core financial primitives:

• The Resolution Agent: Fetches configured oracle data, produces resolution payloads, and records the cryptographic evidence needed for future audits.
• The Settlement Agent: Watches for ready markets, claims a secure task lease, and triggers settlement instantly.
• The DCA Agent: Manages recurring investment strategies by autonomously executing asset purchases at set intervals based on ledger-defined schedules.
• The Oracle Evidence Agent: Stores source metadata such as timestamps, prices, confidence values, data hashes, signatures, and resolver identity.
• The Activity Session Settlement Agent: Batches signed offchain activity receipts and submits settlement or evidence commitments to Canton.
• The Dispute Agent: Detects stale rounds, missing data, or conflicting inputs, automatically moving workflows into manual review before they break.

The Bottom Line

The ledger defines the rules. Agents do the bounded work. Canton records the truth.

The Canton Agent Automation Kit gives builders a clear, safe path to scale their applications without reinventing the security model from scratch. We are making autonomous finance easy to build, safe to operate, and ready for production.