Home
JD.com Launches A2P2 Protocol, First Smart Agent Autonomous Payment Standard with Levels L0 to L5

JD.com has launched the Agent Autonomous Payment Protocol (A2P2), the first systematic protocol in China specifically designed for autonomous payments by intelligent agents. This development marks a pivotal transition for AI payments from conceptual exploration to an established standard.
The protocol outlines six levels of autonomous payment capability for intelligent agents, from L0 to L5. L0 requires complete user authorization for each transaction, while L5 grants agents full autonomy to generate and execute payment decisions. JD.com places significant emphasis on the intermediate levels, L3 and L4. At L3, agents may initiate payment requests within specific tasks, but the system determines approval based on user-defined boundaries. L4 extends this authorization, enabling agents to independently complete payments provided the amount, scenario, and user rules meet pre-set criteria.
In terms of security, the A2P2 protocol introduces an Agent Runtime Identity (ARI) mechanism, which synchronously binds the real user, agent identity, and runtime environment at the point of payment. It performs real-time verification to ensure traceable accountability, unique execution entities, and trustworthy runtime conditions, thereby reducing risks of identity forgery and unauthorized access from the outset.
The protocol also features a fund carrier isolation structure, completely separating agent operations from the user's main account. The system assigns each agent a dedicated restricted account with strict limits on maximum transaction value, usage scenarios, validity period, and payees. Even if an agent is compromised, it cannot exceed its permission or access the main account. Users can monitor the fund usage of each agent in real time and revoke authorization at any time.
For post-transaction settlement and auditing, A2P2 binds payment results with task delegation vouchers, ARI identity information, execution decisions, and tokens to build a comprehensive evidence chain. It introduces an evidence chain mechanism to centrally record key facts, solving the problem of scattered payment information and incomplete audits found in traditional systems, thus delivering verifiable, traceable, and auditable AI transactions.
Industry experts widely believe that the release of A2P2 signals the evolution of intelligent agent payments from product-level features to infrastructure standards, potentially becoming one of the foundational protocols for the future commercialization of AI.
Related article
Anthropic Launches Claude Haiku 5.5, Cutting Small Model Costs by 75% and Letting Users Adjust Inference Intensity
On October 8th, Anthropic introduced its most efficient and capable model to date: Claude Haiku 5.5. Positioned as the fastest and most capable small model in Anthropic’s lineup, it is engineered for high-frequency, cost-sensitive applications such a
Claude Unveils First Full Ultraviolet Sky Map, Charting 119 Million Stars
With the powerful assistance of Claude, Anthropic’s large language model, astrophysicists have successfully generated the first complete ultraviolet all-sky map in history. This vast star chart, featuring 119 million stars, was largely completed by C
How AWS and NVIDIA Are Powering Agentic and Physical AI
NVIDIA CEO Jensen Huang (Credit: Tomohiro Ohsumi/Getty Images)Jensen Huang’s NVIDIA and AWS scale their global partnership, deploying 2M GPUs to power next-generation agentic and physical AI enterprise infrastructureAWS and NVIDIA have announced a ma
Related Special Topic Recommendations
Comments (0)
0/500

JD.com has launched the Agent Autonomous Payment Protocol (A2P2), the first systematic protocol in China specifically designed for autonomous payments by intelligent agents. This development marks a pivotal transition for AI payments from conceptual exploration to an established standard.
The protocol outlines six levels of autonomous payment capability for intelligent agents, from L0 to L5. L0 requires complete user authorization for each transaction, while L5 grants agents full autonomy to generate and execute payment decisions. JD.com places significant emphasis on the intermediate levels, L3 and L4. At L3, agents may initiate payment requests within specific tasks, but the system determines approval based on user-defined boundaries. L4 extends this authorization, enabling agents to independently complete payments provided the amount, scenario, and user rules meet pre-set criteria.
In terms of security, the A2P2 protocol introduces an Agent Runtime Identity (ARI) mechanism, which synchronously binds the real user, agent identity, and runtime environment at the point of payment. It performs real-time verification to ensure traceable accountability, unique execution entities, and trustworthy runtime conditions, thereby reducing risks of identity forgery and unauthorized access from the outset.
The protocol also features a fund carrier isolation structure, completely separating agent operations from the user's main account. The system assigns each agent a dedicated restricted account with strict limits on maximum transaction value, usage scenarios, validity period, and payees. Even if an agent is compromised, it cannot exceed its permission or access the main account. Users can monitor the fund usage of each agent in real time and revoke authorization at any time.
For post-transaction settlement and auditing, A2P2 binds payment results with task delegation vouchers, ARI identity information, execution decisions, and tokens to build a comprehensive evidence chain. It introduces an evidence chain mechanism to centrally record key facts, solving the problem of scattered payment information and incomplete audits found in traditional systems, thus delivering verifiable, traceable, and auditable AI transactions.
Industry experts widely believe that the release of A2P2 signals the evolution of intelligent agent payments from product-level features to infrastructure standards, potentially becoming one of the foundational protocols for the future commercialization of AI.
Anthropic Launches Claude Haiku 5.5, Cutting Small Model Costs by 75% and Letting Users Adjust Inference Intensity
On October 8th, Anthropic introduced its most efficient and capable model to date: Claude Haiku 5.5. Positioned as the fastest and most capable small model in Anthropic’s lineup, it is engineered for high-frequency, cost-sensitive applications such a
Claude Unveils First Full Ultraviolet Sky Map, Charting 119 Million Stars
With the powerful assistance of Claude, Anthropic’s large language model, astrophysicists have successfully generated the first complete ultraviolet all-sky map in history. This vast star chart, featuring 119 million stars, was largely completed by C
How AWS and NVIDIA Are Powering Agentic and Physical AI
NVIDIA CEO Jensen Huang (Credit: Tomohiro Ohsumi/Getty Images)Jensen Huang’s NVIDIA and AWS scale their global partnership, deploying 2M GPUs to power next-generation agentic and physical AI enterprise infrastructureAWS and NVIDIA have announced a ma











