[Feature]: Schedule prompts ( after and in recurring intervals ) #1390
Description
Before submitting
- I searched existing issues and did not find a duplicate.
- I am describing a concrete problem or use case, not just a vague idea.
Area
apps/desktop
Problem or use case
Currently, users must wait for a prompt to finish generating before they can send the next one. Additionally, there is no native way to run recurring prompts (e.g., daily summaries or periodic data checks) without manual intervention or external API scripts.
Proposed solution
ntroduce a Scheduling Toggle in the bottom input bar (adjacent to the "Full Access" settings). This toggle would allow users to define when and how often a prompt should execute.
🛠 User Interface (UI)
Toggle Icon: A clock or calendar icon next to the prompt input field.
Dropdown Options: * Immediate: Standard execution (default).
On Completion: Queue the prompt to run automatically once the current active task finishes.
Recurring: Set an interval (e.g., every 4, 6, or 12 hours) to re-run the prompt.
⚙️ Functional Requirements
- Sequential Queuing ("Run on Completion")
Behavior: If the model is busy, the scheduled prompt is held in a "Pending" state.
Trigger: The moment the active process hits a stop or finished status, the queued prompt is injected into the chat.
Goal: To allow users to "stack" complex workflows without sitting at their desk.
- Recurring Intervals
Behavior: The prompt persists in the background and triggers based on a timer.
Customization: Users can select predefined increments (4h, 6h, 12h, 24h) or input a custom cron-style schedule.
Persistence: The schedule should remain active even if the tab is closed (server-side handling).
Why this matters
ho benefits, and what outcome does this unlock?
Who benefits: * Power Users & Developers: Those running multi-step workflows (e.g., "Code -> Test -> Document") who currently have to wait for each "lap" to finish.
Data Analysts & Researchers: Users who need to monitor shifting data or news throughout the day without manual check-ins.
Busy Professionals: People who want to "set and forget" a prompt before a meeting or a commute and return to a finished result.
Outcome: * Unlocks true asynchronous productivity. It transforms the AI from a synchronous chat tool into a background "agent."
Reduces "Context Switching" fatigue. Users don't have to stay tethered to the tab to hit "Send" the moment the previous generation ends.
Smallest useful scope
The Minimum Viable Product (MVP) would be a simple "Queue Next" button.
Functionality: If a prompt is currently generating, the "Send" button for the next prompt changes to "Queue."
Behavior: Clicking "Queue" simply holds that prompt and sends it automatically the millisecond the current response reaches its finish_reason.
Omission: This version ignores recurring time-based schedules (the 4-hour/6-hour loops) and focuses strictly on sequential execution to solve the immediate bottleneck of waiting for the AI to finish speaking.
Alternatives considered
No response
Risks or tradeoffs
No response
Examples or references
No response
Contribution
- I would be open to helping implement this.