Axelar has launched AgentFlux, an open source framework designed to run artificial intelligence agents locally while keeping private keys, trading strategies and customer data in the cloud – a pitch aimed squarely at institutions exploring on-chain finance but wary of privacy risks.
AgentFlux, developed by Interop Labs, the team behind the Axelar network, allows financial firms to deploy “agent” automation without sending sensitive information to external infrastructure, the company announced Thursday.
The framework solves one of the biggest frictions in AI-driven crypto operations: tool invocation.
Today, most agents rely on cloud-based models to decide which blockchain tools to call and how to structure transactions, which can inadvertently expose information the agency seeks to protect. AgentFlux splits these tasks into two smaller, specialized models—one for selecting the right tool and another for generating the parameters to execute. According to the team behind Axelar, this setup improved the accuracy of tool calls by 46% in benchmark tests, bringing the local model closer to the performance of large cloud systems.
Axelar co-founder Sergey Gorbunov said in an interview with CoinDesk that he believes AgentFlux benefits from multiple tasks: “The first is highly sophisticated trading strategists who have incredible exclusivity,” he said. “You know, you never absolutely want to get a public cloud model that enforces policies like you do. If you upload them, you essentially like to violate them. So that’s one of them.”
“The second thing I do see is that blockchain is also being used heavily for analytics. For example, you are doing tax reporting or investigating who performed which transactions. Or you might have a specific fingerprint and you anticipate malicious activity in a few accounts. Maybe you want to upload that fingerprint to a cloud model or an AI system and ask it to find all related transactions, all these suspicious behaviors and patterns for that account. So you can completely keep those types of activities private. And run them privately,” Gorbunov added.
AgentFlux also feeds into Axelar’s broader multi-chain strategy. The team positions Axelar as a “gateway to on-chain finance,” providing institutions with the infrastructure to move assets and data across blockchains from a single integration point. AgentFlux enables a single AI agent to see risk, assess risk, and transact across multiple ecosystems—a capability the company believes is critical for institutional adoption.
Read more: Canary Capital files documents to launch ETF tracking cross-chain protocol Axelar
