On November 5, the Kimi Open Platform officially launched Kimi Hosted Agents, offering the Harness that supports Agent operation as a cloud service (Harness as a Service). Developers can get a 24/7 hosted environment without building their own runtime framework, with support for long-horizon task execution and multi-agent collaboration. Agents can be configured on demand to build two types of applications: integrated into enterprise internal systems for monitoring and operations response, or embedded into external-facing products to provide in-app investment research assistants and customer service assistants.
The platform provides isolated sandboxes and network access permission controls, allowing users to set operational boundaries and terminate sessions at any time; execution traces are available for review, audit export is supported, and billing rules and usage reports are provided.
The hosted model targets the barriers to Agent adoption. Enterprises often need months to build a Harness foundation, covering the code execution environment, context and session management, tool invocation, permission control, and task interruption recovery, and they still need continuous maintenance after launch. Now Kimi handles the maintenance and upgrades of the underlying environment, which can shorten the cycle from prototype validation to application launch.
The product includes 140+ Skills and 30+ plugins, covering web search, professional data queries, office document processing, and content generation. The standalone security component Kimi Inference Hook checks task inputs and tool return content before every model inference. Security rules are defined by the enterprise, while allow or deny decisions are executed by the Hook, intercepting private and sensitive data before inference. This component is currently available by invitation only to verified enterprise users. The documentation site also provides architecture concepts, configuration tutorials, and cost estimates. The first batch includes eight financial Agent templates, covering investment research modeling, deep research, morning report generation, IPO tracking, and holdings monitoring.
Turning Agent runtime from months of engineering investment into an on-demand cloud service is another step by large model vendors toward extending into productivity deployment.
