Aurora Mobile Limited (NASDAQ: JG) announced that its AI agent platform, GPTBots.ai, has released LoopAgent, an execution engine designed to enable AI agents to autonomously handle complex, multi-step tasks across business systems.

According to the company, LoopAgent is built in-house by the GPTBots team rather than relying on open-source orchestration frameworks. The company says this gives enterprises direct control over the execution pipeline, including security sandboxing, cost guardrails, and step-level audit trails.

Key features included in the release are sandboxed code execution in an isolated Bash environment, lazy-loaded agent skills to reduce token consumption, versioned system identity prompts with diff comparison for compliance and debugging, and automated context summaries for human handoff. The engine also supports real-time retrieval from enterprise knowledge bases.

The company outlined several intended use cases, including insurance claims processing, internal operations automation, complex customer support, and hybrid AI-human service workflows that connect to its LiveDesk and LiveChat products.

"Most agent platforms can handle a conversation. LoopAgent was built to handle a job — start to finish, across systems, with every step traceable," said Chris Lo, Founder and CEO of GPTBots.ai. "The gap between a compelling demo and production-grade AI has always been about execution reliability, not model quality. LoopAgent closes that gap."

LoopAgent is available now on GPTBots.ai. Aurora Mobile, founded in 2011 and headquartered in Hong Kong, provides customer engagement and marketing technology services. The information in this article is based on a press release statement from the company.