- New feature gives agents cross-session long-term memory for personalized responses
- Available free for a limited time via API or OpenClaw plugin
Alibaba Cloud’s “Bailian” Model Studio, the company’s one-stop enterprise AI platform, launched a memory library function on April 9 that equips AI agents with long-term memory across sessions, enabling more personalized user interactions.
The feature is being offered free for a limited period and can be accessed via API or installed in one click through products like OpenClaw.
Most current AI agents lack long-term memory, which can lead to context loss in extended conversations or multi-session tasks, forcing users to repeatedly explain preferences and instructions.
Alibaba Cloud Model Studio’s memory library addresses this with a four-part system—extract, store, retrieve, and inject—allowing agents to automatically capture key information from interactions, retrieve relevant memories on demand, and integrate them into ongoing conversations.
According to the cloud service provider, users can configure and deploy the service in as little as three minutes via the OpenClaw plugin.
Enterprise clients can also call the memory functionality in other agent applications through API, providing configurable, traceable, and evolving long-term memory capabilities.
The system is already applied across education, customer service, emotional companion apps, and smart hardware, supporting enterprise-grade agent memory infrastructure.
Model Studio reported record growth in AI agent usage from January to March 2026, without specifying the growth rates.
As reference, Alibaba disclosed in its fiscal 2026 third-quarter results that Model Studio API token consumption in March 2026 had increased sixfold compared with December 2025.
