> For the complete documentation index, see [llms.txt](https://docs.kinwall.family/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.kinwall.family/self-hosting-reference/cloudflare.md).

# Cloudflare specifics

For deployment steps, see [Deploy to Cloudflare Workers](/getting-started/deploy-cloudflare.md). This page covers how Kinwall behaves on Workers.

## Free-tier limits

| Limit    | Free tier                                                  | How Kinwall fits                                                                                        |
| -------- | ---------------------------------------------------------- | ------------------------------------------------------------------------------------------------------- |
| Requests | 100,000 / day                                              | A wall display polling `/api/rev` every 15 s uses about 5,800/day. Phones add more while open.          |
| CPU      | 10 ms per request or cron run (network wait doesn't count) | Google, Microsoft and CalDAV sync in 31-day slices, and unchanged ICS feeds are skipped by fingerprint. |
| D1       | generous free storage and reads                            | One small database.                                                                                     |

If a very large ICS feed hits the CPU limit, move to **Workers Paid** ($5/month, 30 s CPU). **Sync now** does a full-window sync in one request, so it's the first thing to hit the limit on a huge feed, even when the cron copes.

## Cron

```toml
[triggers]
crons = ["*/5 * * * *"]
```

Each run does both jobs:

* **Sync**: calendars older than `SYNC_INTERVAL_MINUTES` (default 10), stalest first, stopping after about 20 s.
* **Notifications**: reminders, daily summaries and chore nudges. That's why the cron runs every 5 minutes: reminders shouldn't wait for the slower sync interval.

## Placement

```toml
[placement]
mode = "smart"
```

Smart placement runs the Worker near its D1 database rather than near the visitor. A request that makes several database round trips stays fast. Create the D1 database in the location nearest your family (the setup script guesses it from your timezone). An auto-created database lands near Cloudflare's build servers, and every query then crosses that distance.

## Routing

Static assets (the UI) are served by Cloudflare's asset binding with single-page-app fallback. Only these paths reach the Worker: `/api/*`, `/docs`, `/openapi.json`, `/mcp`, `/oauth/*` and `/.well-known/*`.

## Logs

Worker observability is on. When `ADMIN_API_KEY` isn't set, the first-run setup code appears in the Worker's logs (dashboard, or `npx --prefix server wrangler tail`) on the first visit.

## Custom domains

Set them through `CUSTOM_DOMAIN` / the setup script, not only in the dashboard. Each deploy applies `wrangler.toml` and would remove a dashboard-only domain.

## Extra protection

Cloudflare Access in front of the UI (free for up to 50 users), with a service-token bypass for `/api/*`, adds a second login layer.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.kinwall.family/self-hosting-reference/cloudflare.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.