Luffa adds accountable AI agents to encrypted social platform

By AI, Created 11:10 AM UTC, May 20, 2026, /AGP/ – Luffa Technology Network Company Limited says it has launched accountable AI agent capabilities inside its encrypted, decentralized social platform, built on Endless Protocol. The move is meant to give AI agents identity, permissions, revocation, and auditability as social apps and agent-native services converge.

Why it matters: - Luffa is positioning AI agents as governed digital actors, not hidden automation inside a messaging app. - The framework is meant to give users and developers clearer control over what an AI agent is, what it can do, and who authorized it. - The company says the architecture is designed for the emerging AI Agent Economy, where encrypted communication, identity, permissions, and execution need to work together.

What happened: - Luffa Technology Network Company Limited announced accountable AI agent capabilities inside its encrypted, decentralized social platform. - The launch is built on Endless Protocol. - Luffa says the update extends decentralized identity from human users to AI agents. - The platform is available on iOS and Android. - A More information link is available in the company announcement: the company’s announcement

The details: - Each Luffa agent can have a verifiable identity layer. - That identity layer is designed to show what the agent is, who or what it represents, what permissions it has, who authorized it, what actions it can perform, and how access can be changed or revoked. - Luffa says this creates Agent DID, or decentralized identity for autonomous AI systems. - The framework is built around cryptographic data isolation, so private user messages are not accessible to AI agents by default. - Luffa requires clear disclosure when an agent or bot is non-human. - Users control which agents can interact with them, what those agents can access, and when permissions are removed. - Luffa says private conversations are not treated as model-training material. - The architecture is designed to move AI governance away from unilateral platform control and toward protocol-based accountability.

Between the lines: - Luffa is trying to solve a trust problem that will grow as AI agents become more autonomous and more embedded in everyday digital life. - The company is also drawing a line against hidden AI personas, undisclosed automation, and platform-owned training use of private conversations. - This framing puts Luffa in a broader race to define the rules for agent identity before centralized platforms set the standards.

What’s next: - Luffa plans to support use cases including translation agents, customer service and community management bots, agent-based service accounts, SuperBox mini-app integrations, secure agent-to-user flows, and AI-assisted execution across social, commercial, and decentralized applications. - The company says it aims to expand into a broader agent-native ecosystem where users, communities, brands, developers, services, and AI agents operate in one trusted environment. - Luffa says its next phase is about building a trust infrastructure for human-agent interaction, not just adding AI features to social messaging.

The bottom line: - Luffa is betting that the next wave of social software will depend on verifiable identity, permission control, and encrypted execution for AI agents, not just faster chat or smarter bots.