AI Agents to Perform Cross-Chain Operations via deBridge

The team behind decentralized cross-chain protocol deBridge announced the launch of deBridge MCP, an open-source server plugin based on the Model Context Protocol (MCP). It integrates with developer tools and AI agents, granting them access to direct on-chain transaction execution.

deBridge MCP enables transactions between EVM networks and within the Solana blockchain through a unified interface. Users retain control over their assets at every stage, while the execution process is abstracted from technical complexities. The solution supports optimal route calculation, intra-network swaps, cross-chain asset transfers, and automated transaction routing.

According to the press release, the solution implements a deterministic execution model, meaning fees are charged according to predefined parameters and routing is optimized with MEV considerations to enhance reliability. The developers state that this approach reduces risks associated with automating cross-chain operations and eliminates common challenges of working across multiple blockchains.

Among the potential use cases of deBridge MCP, the team highlighted:

1. Vibe Trading. A trading model in which the user specifies the desired outcome (intent) without defining the exact steps to achieve it. The AI agent independently plans and executes all necessary on-chain actions.
2. Algorithmic portfolio management. AI agents can continuously analyze asset allocation, assess risk concentration in specific networks or tokens, and automatically rebalance portfolios. If excessive exposure in a particular network is detected, the agent calculates optimal swap and bridge routes and executes transactions without requiring custom scripts.
3. On-chain request execution with immediate settlement. The agent can determine the most cost-efficient route for transferring assets between networks, compare fees, evaluate slippage, and execute the transaction after user confirmation. The system supports retrieving token lists, price oracle data, and network information with subsequent aggregation of results.
4. Orchestration of multi-step operations under a single intent. For example, when consolidating assets from multiple networks into a single token on a selected network for staking, the agent automatically combines bridge operations, swaps, and DeFi interactions into a single atomic process based on deBridge’s deterministic execution layer.

Recently, Coinbase introduced Agentic Wallets, wallets that enable autonomous AI agents to manage assets, execute transactions, and pay for services without human intervention. More about autonomous AI agents can be found in CP Media’s special report.