Vercel Schedules
A dynamic task scheduling product from Vercel that supports static cron and runtime dynamic scheduling. It allows AI agents to create, modify and cancel tasks through an API without redeploying.
Timeline (last 90 days)
Vercel Schedules is opened to both Agents and applications, with scheduling authority delegated from operations configuration down to the application logic itself.
Dynamically creating tasks at runtime is important for Agents because it leaves the decision of "when to check next" within the application logic while handing "wake up at the scheduled time" to the platform for execution.
Handing scheduling authority to the application itself follows the trend of Agents moving from "passive response" to "proactive execution".
Vercel Schedules is currently in public beta, and its feature boundaries may be adjusted based on feedback.
The specific quotas, limits and billing methods of Vercel Schedules have not yet been detailed in official announcements.
For applications using the Nitro 3 server framework, Vercel Schedules provides an experimental integration that allows periodic tasks in existing scheduledTasks configurations to run as private functions.
Vercel Schedules is currently in Public Beta.
Vercel Schedules is built on top of Vercel Cron Jobs' static approach, allowing users and AI agents to decide when tasks execute while the application is running, without redeploying.
With Vercel Schedules, agent scenarios no longer require a separate database and polling jobs; agents can directly create scheduling records requesting content and execution times.
The eve agent framework directly uses Vercel Schedules to provide dynamic scheduling capabilities for production agents hosted on Vercel.
4 more timeline events
All Facts (14)
Vercel Schedules is opened to both Agents and applications, with scheduling authority delegated from operations configuration down to the application logic itself.
50%UnverifiedDynamically creating tasks at runtime is important for Agents because it leaves the decision of "when to check next" within the application logic while handing "wake up at the scheduled time" to the platform for execution.
50%UnverifiedHanding scheduling authority to the application itself follows the trend of Agents moving from "passive response" to "proactive execution".
50%UnverifiedVercel Schedules is currently in public beta, and its feature boundaries may be adjusted based on feedback.
50%UnverifiedThe specific quotas, limits and billing methods of Vercel Schedules have not yet been detailed in official announcements.
50%UnverifiedFor applications using the Nitro 3 server framework, Vercel Schedules provides an experimental integration that allows periodic tasks in existing scheduledTasks configurations to run as private functions.
50%UnverifiedVercel Schedules is currently in Public Beta.
50%UnverifiedVercel Schedules is built on top of Vercel Cron Jobs' static approach, allowing users and AI agents to decide when tasks execute while the application is running, without redeploying.
50%UnverifiedWith Vercel Schedules, agent scenarios no longer require a separate database and polling jobs; agents can directly create scheduling records requesting content and execution times.
50%UnverifiedThe eve agent framework directly uses Vercel Schedules to provide dynamic scheduling capabilities for production agents hosted on Vercel.
50%UnverifiedVercel Schedules achieves dynamic scheduling through underlying Queues and function invocation mechanisms, with scheduling information stored on the platform side rather than in the codebase.
50%UnverifiedIn the Vercel Schedules example, the agent creates a schedule through AI SDK tools, converting natural language instructions into structured schedule-creation requests.
50%UnverifiedVercel Schedules emphasizes that permission checks occur at task execution time rather than creation time, ensuring that execution is still correctly validated after a user's permissions change.
50%UnverifiedVercel Schedules supports testing scheduled tasks in Preview deployments and custom environments, manually triggering them and viewing execution history via the "Run Schedule" feature.
50%