# Scheduled deliveries

> Deliver a dashboard or chart on a recurring schedule to email, Slack, or a webhook — optionally only when a metric crosses an alert threshold.

*[View this page in the Flax docs](https://flax-analytics.com/docs/automation-ai/scheduled-deliveries)*

A scheduled delivery sends a dashboard or chart to people on a recurring cadence,
so a report lands without anyone opening Flax. Optionally, attach an **alert** so a
delivery only fires when a metric crosses a threshold. Set up personal deliveries
from a dashboard's own delivery tab; admins oversee all of them at
`/app/settings/automation`.

## What you can deliver

Deliveries carry either a **dashboard** or a **chart** and send it over one of three
channels:

| Channel | Delivery |
| --- | --- |
| **Email** | The dashboard rendered inline, with a PDF attached and a link back to Flax. |
| **Slack** | Posts to a Slack incoming webhook. |
| **Webhook** | Signed HTTP POST to your endpoint. |

Recipients are set by the delivery's **audience**: specific emails (custom), a
team, or the whole organization.

## Cadence

Schedules run on a structured cadence, interpreted in **UTC** — not raw cron:

| Frequency | Fields |
| --- | --- |
| **Hourly** | Minute (0–59). |
| **Daily** | Hour and minute. |
| **Weekly** | Weekday, hour, and minute. |

## Alerts

Attach an alert so a delivery fires only when a metric moves. An alert names a
model and measure and a comparison against a threshold:

| Operator | Meaning |
| --- | --- |
| `gt` / `gte` | greater than / or equal |
| `lt` / `lte` | less than / or equal |
| `eq` / `neq` | equal / not equal |

When it runs, Flax evaluates a one-row query for the measure and only delivers if
the comparison holds.

## When a report can't be rendered

An email delivery renders the dashboard as its owner, under exactly their access. If
that render fails — a slow warehouse, a timeout, a temporary outage — Flax **does not
send a partial email**. A report that arrives without its preview and PDF looks like a
report saying nothing happened, so instead the delivery is retried after 5, 15 and 45
minutes. If it still can't be built after that, nothing is sent for that run and the
next scheduled run proceeds normally. A retry that would overrun the next scheduled run
is skipped in favour of the fresher report.

Some deliveries are text notices by design, and those are unaffected: **alerts** (which
report a threshold, not a dashboard), resources that have no printable view, and
organizations without the Export capability. They send a short notice with a link.

Self-hosting: set `FLAX_ERROR_EMAIL` to receive a batched digest whenever a delivery
gives up, along with any other application errors.

## Managing deliveries

Create, pause, and delete deliveries at `/app/settings/automation`. Org admins see
every delivery in the organization; team leads see their teams'. Each row shows the
owner, resource, schedule, channel, next run, and an active/pause toggle. Personal
"specific emails" schedules are created and edited from the dashboard they deliver.

## Related

- [Actions](/docs/automation-ai/actions) — schedule and threshold triggers that run effects.
- [Exporting charts](/docs/visualizations/exporting) — server-side rendering for downloads and embeds.