FTP Tech launches MCP server to link AI agents to Canton Network wallets

Connecting AI to On-Chain Data

FTP Tech has launched a new server that enables popular AI assistants, such as ChatGPT and Claude, to interact directly with the Canton Network. This is made possible through the Model Context Protocol, or MCP. MCP is an open standard designed to let chatbots reach external tools and data without requiring a unique custom integration for every single service.

By using this standard, the AI does not need to be rebuilt to understand a specific wallet or network. Instead, the MCP server acts as a bridge. It allows the AI to 'see' and interact with the infrastructure provided by the Canton Network, making it easier for developers to build agents that can perform complex tasks.

As Cryptonews.net reported, FTP Tech is a node-as-a-service provider and a core contributor to the Canton ecosystem. They have been building in the Web3 space since 2021, focusing on the underlying infrastructure that allows different software systems to talk to one another and execute commands.

How Agentic Commerce Works

The primary goal of this technology is to facilitate agentic commerce. This refers to a system where automated AI agents can act on behalf of individual users to conduct business. In this specific setup, the AI can access on-chain wallets to move funds or interact with financial instruments directly on the blockchain.

The demo of this technology shows Canton Coin (CC) being used by these agents. While CC is typically used to pay network usage fees, this integration shows it also serving as a payment and trading currency. The agents can use CC to buy various tokenized assets, such as bonds, equities, exchange-traded funds, and money-market instruments.

The Canton Network itself is a privacy-focused public Layer 1 network designed for institutional finance. It supports stablecoin settlements and various tokenized assets. By linking these assets to an AI agent via an MCP server, the system creates a path for automated trading and asset management that was previously difficult to implement.

Safety and Human Oversight

Giving AI agents direct access to financial wallets is still a very early stage of development. To address the risks of an AI making unauthorized purchases, FTP Tech has included spending caps in its design. These limits are intended to keep a human in control of the overall budget while allowing the agent to operate independently within those boundaries.

This balance is crucial for users who want the convenience of automation without the risk of an agent spending unlimited funds. The design ensures that the agent can only perform actions that fall within the pre-approved limits set by the human owner. It creates a controlled environment for the AI to execute trades or move assets.

For users who eventually adopt these tools, the security of the wallet is the primary concern. Experts suggest that anyone considering such tools should carefully check the cap settings and the specific permissions granted to the agent before they fund a wallet. This ensures the AI has the power to do what is needed but not more.

The Role of Canton Coin

Canton Coin (CC) serves as the native utility token for the Canton Network and plays a major role in the network's economics. Institutions currently pay usage fees in CC rather than in traditional dollars. A portion of every fee paid in CC is permanently burned, which helps manage the token's supply and utility within the ecosystem.

The new FTP Tech integration highlights a different, more active use for the token. Instead of just being a fee-paying mechanism, it becomes a functional currency for automated agents. This allows the AI to interact with the broader financial world by using CC to settle transactions and acquire various types of tokenized securities.

This dual role for CC—as both a network fee and a trading currency—is central to the ecosystem's growth. It provides a unified way for both human institutions and automated agents to participate in the network. As more agents are deployed, the demand for CC as a trading currency for automated commerce may increase.

What Comes Next for AI Agents

While this integration is a significant step forward, many details remain unconfirmed. The specific assets that can be traded beyond those shown in the demo have not been independently verified. This means the current capabilities are a starting point for what is possible when AI meets institutional finance infrastructure.

The move toward agentic commerce suggests a future where AI assistants are no longer just for writing emails or summarizing text. They will become active participants in the economy, capable of managing portfolios and executing trades. This requires a robust set of standards like MCP to ensure these agents can connect to the tools they need to function.

As the technology matures, we can expect to see more node-as-a-service providers offering similar integrations. The focus will likely shift toward refining the safety protocols and spending caps to make it even easier for non-technical users to deploy agents safely. For now, the focus remains on building the foundational links between large language models and on-chain financial systems.