Your AI tasks now schedule themselves. Kimi Work ships a cron engine.

Moonshot AI's Kimi platform now includes a scheduled automation feature called Kimi Work. It runs LLM agent calls, Python scripts, and similar tasks on a cron timer, without requiring a constantly connected local machine.

The core idea is straightforward: you define a recurring task and a schedule, and Kimi Work executes it in the background. Early-morning briefings, midnight dataset passes, regularly scheduled data pulls, anything that currently forces someone to start a job and wait for it can be set to run unattended. For similar hands-off automation, see Kimi Claw's one-click cloud deployment.

The feature is built into Kimi's existing cloud infrastructure, so users do not need to provision a server, install a scheduler, or keep a laptop open. It is the same kind of promise behind Kimi Claw, which lets users deploy browser automation agents without local setup. Kimi Work extends the same philosophy to scripted and LLM-driven tasks.

On the surface, this is a convenience feature. Underneath, it points to a shift in how AI platforms think about workflow. Most API-based agents today require a server-side scheduler or an always-on client. Kimi Work folds scheduling into the product layer, making it accessible to users who do not want to write cron files or manage infrastructure. Competitors like Groq are also moving toward agent operating systems, albeit from a different angle.

The question that matters is how much users actually trust these automated runs. An unattended LLM call that fetches stale data or produces a hallucinated report is worse than no report at all. Kimi Work itself is a scheduling engine, the quality of the output still depends on the model and the prompt design behind each task. The problem of evaluating AI outputs at scale is not trivial, as recent research on agent self-evaluation shows.

Who is it for?

Kimi Work is most useful for teams with regular, well-defined reporting needs: a daily market summary, a nightly code analysis run, a weekly data aggregation. It also suits solo developers who want to decouple batch processing from their own work hours. The growing popularity of Kimi's models on platforms like Opencode suggests a user base ready for deeper integration.

Users who need event-triggered workflows (run task X when file Y changes) will still need an external orchestrator. Kimi Work is purely schedule-driven, not event-driven.

Moonshot AI has not released pricing details for Kimi Work beyond saying it is included in the existing Kimi platform tiers. The feature is available now to all users. The launch comes amid high demand, as recent subscription pauses indicated earlier.