> ## Documentation Index
> Fetch the complete documentation index at: https://keystroke.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.
# Schedules
> Run a workflow on a recurring schedule.
A schedule trigger runs a [workflow](/docs/learn/workflows/overview) or [agent](/docs/learn/agents/overview) on a recurring cron schedule. Use it for anything time-driven: a morning digest, an hourly sync, a nightly cleanup.
## Example requests
Ask your coding agent what should happen and how often. It can set up the schedule and attach it.
> "Every weekday at 8am, run the morning briefing workflow and post it to Slack."
> "On the first of each month, have the FP\&A agent review recurring charges and flag anything unused."
> "Every night, sync new rows from the form submissions table in Postgres into the reporting spreadsheet."
## Define a schedule
Use `defineCronSource` with a `slug`, `name`, `description`, and a cron `schedule`, then attach a target.
```ts src/triggers/morning-check.ts theme={null}
import { defineCronSource } from "@keystrokehq/keystroke/trigger";
import workflow from "../workflows/morning-check";
export default defineCronSource({
slug: "morning-check",
name: "Morning check",
description: "Runs the morning check workflow every day at 09:00.",
schedule: "0 9 * * *",
}).attach({ workflow });
```
| Option | Required | What it does |
| ------------- | -------- | --------------------------------------------------------------------------------------------------------- |
| `slug` | Yes | Stable trigger slug (used in the [attachment id](/docs/learn/triggers/overview#attachment-id) and run history) |
| `name` | Yes | Human-readable label shown in the platform |
| `description` | Yes | What the schedule does, shown in the platform |
| `schedule` | Yes | A cron expression for when the trigger fires |
| `timezone` | No | IANA timezone for schedule wall-clock fields (e.g. `America/New_York`). When omitted, UTC |
| `payload` | No | Static JSON passed to attachments on each tick (defaults to `{}`) |
## The cron schedule
The `schedule` is a standard cron expression. The common five-field form is `minute hour day-of-month month day-of-week`:
```ts theme={null}
schedule: "0 9 * * *" // every day at 09:00
schedule: "*/15 * * * *" // every 15 minutes
schedule: "0 0 * * 1" // every Monday at midnight
```
Named months and weekdays (`MON`, `JAN`) and shorthand nicknames (`@daily`, `@hourly`) are also accepted.
Pass an optional IANA `timezone` so the cron fields are wall-clock times in that zone (DST-aware). When omitted, schedules evaluate in UTC:
```ts theme={null}
defineCronSource({
slug: "morning-check",
name: "Morning check",
description: "Runs the morning check workflow every day at 09:00 Eastern.",
schedule: "0 9 * * *",
timezone: "America/New_York", // 9:00 AM Eastern every day
}).attach({ workflow });
```
Prefer `timezone` for wall-clock times like "9am" in a specific region. Without it, `"0 9 * * *"` means 09:00 UTC.
## Input / payload
By default a schedule fires with an empty input (`{}`). Pass an optional static `payload` on the source when the workflow needs input — it flows to attachments the same way a webhook body does:
```ts src/triggers/morning-report.ts theme={null}
import { defineCronSource } from "@keystrokehq/keystroke/trigger";
import workflow from "../workflows/morning-report";
export default defineCronSource({
slug: "morning-report",
name: "Morning report",
description: "Runs the daily regional report.",
schedule: "0 9 * * *",
payload: { report: "daily", region: "us" },
}).attach({ workflow });
```
Without a `transform`, the payload must match the workflow's `input` schema (TypeScript enforces this at the `.attach()` call). Use `transform` to reshape the payload per attachment:
```ts theme={null}
.attach({
workflow: euReport,
transform: (payload) => ({ ...payload, region: "eu" }),
})
```
When `payload` is omitted, attachments still receive `{}`, so empty-input workflows keep working:
```ts theme={null}
input: z.object({}),
```
A schedule attached to an agent can use a static `prompt`, or a `prompt(payload)` function that reads the cron payload:
```ts src/triggers/morning-check.ts theme={null}
import { defineCronSource } from "@keystrokehq/keystroke/trigger";
import support from "../agents/support";
export default defineCronSource({
slug: "morning-check",
name: "Morning check",
description: "Asks support to summarize overnight issues.",
schedule: "0 9 * * *",
payload: { focus: "overnight" },
}).attach({
agent: support,
prompt: (payload) =>
`Run the morning check (focus: ${(payload as { focus: string }).focus}) and summarize anything that needs attention.`,
});
```
## Test it while building
You don't have to wait for the clock. After deploy, fire the cron trigger immediately (same static payload as a scheduled tick — platform only):
```bash theme={null}
keystroke triggers invoke morning-check
keystroke triggers invoke morning-check --workflow morning-check # narrow to one attachment
```
Or invoke the target directly while developing:
```bash theme={null}
keystroke workflows run morning-check --input '{}'
keystroke agents prompt support --message "Run the morning check."
```
After deploy, the platform also fires the schedule on time — every tick runs the target unconditionally (cron has no filters). Inspect runs with `keystroke triggers runs list morning-check --workflow ` (or `--agent `).
## Next steps
Run on a schedule, but only when there's new work.
Run on inbound HTTP requests instead of a clock.
Agent prompts, transforms, and filtering.
Sources, attach, and attachment ids.