> ## Documentation Index
> Fetch the complete documentation index at: https://docs.letsfind.work/llms.txt
> Use this file to discover all available pages before exploring further.

# Changes and webhooks

> Keep your copy of the jobs in sync: new, changed and closed jobs since your last call.

`GET /v1/changes` returns jobs that were created, changed or closed, in the order they changed, with a cursor. It takes every filter of `GET /v1/jobs`.

## Keep a copy in sync

1. Call `GET /v1/changes` without `cursor`. You get every open job that matches your filters, as `created`.
2. Save `next`. Call again with `cursor=<next>` whenever you like: every 5 minutes or once a week. You never miss a change and you never need to call at a fixed time.
3. While `more` is `true`, call again at once.

```js theme={null}
let cursor = await load() // null the first time
for (;;) {
  const url = new URL("/v1/changes", process.env.BASE_URL)
  url.searchParams.set("country", "DE")
  url.searchParams.set("limit", "100")
  if (cursor) url.searchParams.set("cursor", cursor)
  const page = await (await fetch(url, { headers: { Authorization: `Bearer ${process.env.API_KEY}` } })).json()
  for (const c of page.changes) {
    if (c.type === "closed") await remove(c.id)
    else await upsert(c.job) // created or changed
  }
  cursor = page.next
  await save(cursor)
  if (!page.more) break
}
```

| Change | What you get |
| - | - |
| `created` | The job, new since your cursor. |
| `changed` | The job as it is now. A job changed several times since your cursor comes once. |
| `closed` | The job id, when we found it closed (`closedAt`) and why (`reason`): `removed` (gone from its board), `not_found` (its page is gone), `expired` (past its valid date) or `too_old`. |

Each change also has its own `cursor`, so you can go on from the middle of a page.

## What it costs

* A `created` or `changed` job costs one credit, unless you already got it this month (from any endpoint, the MCP server or a webhook).
* `closed` events are free.
* Calling twice with the same cursor costs nothing the second time.
* On the free plan, a page stops before the first job past your quota, and `next` points there. Nothing is skipped.

A cursor is good for 30 days. An older one gets `410`: start again without a cursor (jobs you already got this month are free).

Closed events of jobs we no longer store are not filtered: ignore ids you do not have.

## Webhooks

Every paid plan can add up to 5 webhooks on the account page or with `POST /v1/webhooks` (`url`, optional `filters`). From then on, we send the changes that match to your HTTPS URL every 30 minutes, in batches of up to 500, billed like `GET /v1/changes`. The body is a page of changes with the webhook's id:

```json theme={null}
{ "webhook": "wh_...", "changes": [ ... ], "next": "...", "more": false }
```

* Answer with any `2xx` within 10 seconds. We do not follow redirects.
* Each request is signed the [Standard Webhooks](https://www.standardwebhooks.com) way: headers `webhook-id`, `webhook-timestamp` and `webhook-signature` (`v1,` and a base64 HMAC-SHA256 of `id.timestamp.body`, keyed with the base64 part of your `whsec_` secret). Their libraries check it for you.
* A failed batch is sent again later with the same `webhook-id` (30 minutes, then doubling up to a day). After 8 failures in a row the webhook is turned off; add it again to start from that moment.
* A batch may come twice if your answer is lost: drop a repeated `webhook-id`.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.