TermiX has released the technical whitepaper for AACP, or Agent Autonomous Commerce Protocol, an open framework designed to let AI agents hire one another, perform tasks and settle payments through on-chain infrastructure. In its official AACP announcement, TermiX describes a system combining identity, escrow, verification and dispute resolution. The protocol is aimed at giving autonomous agents a structured commercial lifecycle rather than simply providing another payment rail.
The framework separates a commercial interaction into stages covering task funding, execution, verification and settlement. Agents can publish jobs, provide collateral, submit offers and deliver work, while funds remain in escrow until predefined conditions are satisfied. AACP attempts to make agent-to-agent commerce enforceable through programmable rules instead of relying entirely on trust between software operators.
AACP Combines Identity, Escrow and Verification
The open-source AACP specification on GitHub builds its architecture around ERC-8004 for agent identity and reputation and ERC-8183 for job escrow and lifecycle management. It also proposes trusted execution environments and zero-knowledge virtual machines for stronger verification of completed work. The design links an agent’s identity and reputation directly to jobs that can be funded, challenged and settled on-chain.
AACP also introduces persistent staking and slashing. Participants lock collateral that can support multiple jobs, while dishonest behavior can result in penalties. Disputes are designed to move to randomly selected arbitrators whose decisions determine settlement. The economic model is intended to make reputation and capital exposure part of the security system rather than relying solely on software permissions.
TermiX’s existing marketplace already illustrates parts of that model. Its Agent.family marketplace allows agents to establish .agent identities, publish services, accept jobs and use on-chain escrow. Minting an agent identity, however, is only one component of the system and does not by itself grant autonomous commercial authority.
Full AACP Deployment Is Still Ahead
The release should not be interpreted as evidence that every component of AACP is already operating in production. TermiX’s GitHub roadmap places staking and reputation on Base Sepolia in Q2 2026, verification infrastructure on Base mainnet in Q3, dispute resolution in Q4 and full deployment across Base and BSC in Q1 2027. AACP is therefore moving through staged implementation rather than launching as a finished settlement network immediately.
The broader effort also reflects an industry push toward standards that let independent agents work together. The Linux Foundation’s Agentic AI program highlights interoperability, identity, permissions and secure coordination as core requirements for production agent systems, while its Agent2Agent initiative focuses specifically on communication between agents. AACP targets a different layer of that emerging stack: the economic rules governing jobs, collateral, verification and payment.
TermiX has published a detailed architecture and begun building software around it, including agent skills operating against a BSC testnet environment. The critical test will be whether those mechanisms can progress from specification and controlled deployments into reliable agent-to-agent commerce at mainnet scale.
Natalie Pierce tracks the parts of crypto that move fast and rarely wait for everyone to catch up. From South Africa, she covers DeFi, AI crypto, hacks, airdrops, sentiment and emerging narratives, especially when user behavior and protocol risk start to overlap.
Her reporting is built for messy sectors. Natalie looks at incentives, reactions, security concerns, social momentum and early signs of traction without pretending every new trend is already proven. Her voice is clear and accessible, but careful enough for areas where excitement can outrun the facts very quickly.