---
title: "Recurring XML job feeds"
label: "Recurring XML feeds"
description: "Pull jobs from a public XML feed URL on a schedule. AI maps fields on the first run, then Cavuno updates, adds, and optionally expires listings automatically."
publishedAt: 2026-07-09
updatedAt: 2026-10-20
order: 2
status: "published"
faqs:
  - question: Where do I add a recurring feed?
    answer: "Go to Backfill → Feeds, click Add feed, and paste a public XML feed URL."
  - question: Why was my feed URL rejected?
    answer: "Cavuno checks the URL when you add a feed or change its URL, and only saves it if it returns XML. Search results on job sites such as LinkedIn or Indeed are web pages, not feeds. You can use backfill rules to bring in jobs like these automatically. For an ATS job board such as Ashby, Lever, Greenhouse, or Workable, you can add the company under Backfill → Companies."
  - question: How often can the feed run?
    answer: "Every 1, 3, 6, 12, or 24 hours. The default is 6 hours. The first sync starts within a few minutes of adding the feed."
  - question: Can I change the feed URL later?
    answer: "Yes. Open the feed, edit Feed URL, and click Save URL. Cavuno checks the new URL the same way as a new feed, and the next sync fully reparses the new source."
  - question: Will an empty feed delete all my jobs?
    answer: "No. Empty or unparseable feeds never expire listings. Expiry only runs after a job is missing from two consecutive successful runs, and only when Expire missing jobs is enabled."
  - question: Can I cap how many jobs one feed adds?
    answer: "Yes. Set Maximum jobs from this feed when you add the feed, or later from the feed's detail page. It caps how many published jobs from that feed can be live at once, underneath your plan limit and the board-wide Max jobs backfilled through feeds. It only stops new imports; jobs already live stay until they expire."
  - question: What does Needs review mean?
    answer: "The feed paused expiry or further risky changes because a run looked unsafe (for example more than 40% of active feed jobs missing at once). Fix the feed or mapping, then resume from the feed detail page."
---

Keep a board in sync with a partner, ATS, or aggregator export without re-uploading files. Add a public XML feed URL under **Backfill → Feeds**, and Cavuno pulls it on a schedule.

One-shot uploads are covered in [Import jobs from CSV or XML](/docs/import-export/job-import).

## Before you begin

You need a stable, publicly reachable `http` or `https` URL that returns XML or gzip-compressed XML without an interactive login. Confirm that the feed includes a stable application URL or external ID for each job, and decide whether listings missing from two consecutive successful runs should expire.

### URLs that are not feeds

When you add a feed or change its URL, Cavuno fetches the URL once and only saves it if the response is XML. If the URL returns a web page, JSON, or an error, the form tells you why. Each board can run up to 20 of these checks per hour.

- Search results on job sites such as LinkedIn or Indeed are web pages, not feeds. You can [use backfill rules](/docs/backfill/configure-backfill#create-a-rule) to bring in jobs like these automatically
- ATS job-board URLs, such as Ashby, Lever, Greenhouse, or Workable boards, are not XML feeds. You can add the company under **Backfill → Companies** to [import its jobs](/docs/backfill/configure-backfill#adding-a-new-company-to-the-backfill)

## Add a feed

1. Go to **Backfill → Feeds**
2. Click **Add feed**
3. Paste the **Feed URL** (public `http` or `https` XML; gzip is supported)
4. Optionally set a **Feed name** so the list is easy to scan
5. Choose **Check every**: 1, 3, 6, 12, or 24 hours (default 6)
6. Leave **Expire missing jobs** on if jobs that leave the feed should expire after two consecutive misses
7. Optionally set **Maximum jobs from this feed** to cap how many of its jobs can be live at once. Leave it empty for no per-feed limit
8. Click **Add feed**. Cavuno checks the URL first; the first sync starts within a few minutes

You can have up to 10 feeds per board.

## How sync works

### First run

Cavuno fetches the feed and runs the same AI field mapping as one-shot import. Mapping is stored for later runs so you do not re-map every cycle.

### Later runs

- **Unchanged body:** if the feed content has not changed, the run short-circuits without rewriting jobs
- **New or changed jobs:** matched by external ID / reference number when the feed sends one, otherwise by application URL together with the job title, then created or updated in place. Different roles that share one application URL or email (such as a single careers page) import as separate jobs
- **Missing jobs:** when **Expire missing jobs** is on, a job missing for two consecutive successful runs is expired (about 12 hours at the default 6-hour cadence)
- **Rows that can't be imported:** a row with missing required fields, an invalid expiry date, or a company Cavuno can't create is skipped and listed with its reason in the run history. The rest of the feed keeps importing, and a job that row already added to your board stays live

### Safety guards

- An empty or unparseable feed never expires anything
- If more than about 40% of active feed jobs would go missing in one run, the feed moves to **Needs review** and expiry is held back
- If more than about 40% of a run's rows can't be imported, the feed moves to **Needs review**, since the feed or its mapping is likely broken
- A new feed URL is checked before it is saved, and the next run fully reparses the new source

## Feed status

| Status | Meaning |
|---|---|
| First sync pending | Mapping or first pull has not finished yet |
| Active | Running on schedule |
| Failing | Recent pulls failed (network, parse, or similar) |
| Needs review | A safety guard fired; check the feed detail and last error |
| Paused | The feed is disabled in settings |

Open a feed from the list to see last run, next run, interval, expire settings, the per-feed job limit, and recent run history. Interval and expire options save automatically; the job limit saves when you click **Save limit**.

## Check the field mapping

The **Field mapping** section on a feed shows how the feed fills each job field, with example values from the latest run:

- **Required** lists the job title, the company, and how to apply (an application URL or email). Jobs import once all three have a source
- **Optional** lists the other job fields the feed fills, such as location, salary, or a job ID
- Below them, you can expand the job fields the feed doesn't provide and the feed columns Cavuno doesn't use

Some fields are built from part of a column, such as a job ID taken from the text before "|" in a `guid` column. The row names the column and the part used, and the example shows the result.

To change the mapping, click **Edit mapping**. You can point each feed column at a different job field, or remove a built field so the field maps from its column again. Changes apply from the next run.

If the feed's columns change and the mapping no longer fits, imports pause and the feed moves to **Needs review**. Click **Re-detect mapping** to rebuild the mapping from the latest feed; the option is also in the **⋮** menu next to **Edit mapping**. Re-detecting keeps how job IDs are read, so jobs already on your board stay matched.

## Limit how many jobs a feed adds

Each feed can carry its own ceiling: **Maximum jobs from this feed** is the number of published jobs from that feed that can be live at the same time. It sits underneath your plan's active jobs limit and **Max jobs backfilled through feeds**, the board-wide [limit for all feeds](/docs/backfill/configure-backfill#backfill-limit), so the smallest of the three wins.

Use it to keep one busy feed from crowding out the rest of your sourcing. For example, a board with feeds capped at 200 in total can give one partner feed 150 and another 50.

Like the board-wide limit, a feed limit only stops new imports. Jobs already on the board stay until they expire or drop out of the feed, and each run tops the feed back up to its limit as room frees up.

## Manage feeds with the API, MCP, or CLI

Agents and scripts can connect and run feeds without the dashboard. Feeds are Operator API resources under `/job-feeds` in the [API reference](/docs/api): list, create, read, update (URL, name, interval, enabled, expiry, job limit, apply link attribution, and field mapping in one request), delete, re-detect the mapping, and list a feed's runs. The Cavuno MCP server reaches the same operations, and the CLI has a `cavuno job-feeds` command group.

Reading feeds and runs requires the `imports.read` scope. Creating, editing, re-detecting, and deleting require `imports.manage`.

A new feed URL goes through the same check as the dashboard, and counts toward the same 20 checks per hour. The first run detects the mapping, so a new feed reads `pending_mapping` with `mapping: null` until it finishes. Read the feed again to see the detected mapping, then correct it with an update if needed. Re-detecting the mapping returns the feed to that state.

```bash
cavuno job-feeds create --url https://ats.example.com/jobs.xml --interval-hours 12
cavuno job-feeds get <feed-id>
cavuno job-feeds runs <feed-id>
```

## When to use a feed

- Partner or association boards that publish a shared XML feed
- ATS or job-distribution exports on a stable URL
- PPC or affiliate partners that supply tracked application URLs in an XML feed

## Use a feed for PPC monetization

If you have a separate PPC or affiliate agreement, map the provider's tracked destination into the job's application URL. Cavuno preserves that URL when candidates click **Apply**, allowing the provider to perform its own attribution. Cavuno does not create the partner agreement, add missing tracking parameters, decide which clicks qualify, or issue payouts.

Before publishing the feed broadly, open one imported job, follow its application link, and confirm the provider records the test according to your agreement. Compare Cavuno's outbound apply-click data with the provider's accepted-action report; the two can differ because the provider controls validation and attribution.

## Verify the feed

Open the feed detail after the first sync. Confirm the status is **Active**, review the imported and skipped counts, and open representative jobs to check the title, company, location, description, and application URL. If you enabled expiry, remove a test job from the source only after validating the mapping and confirm Cavuno expires it after two consecutive successful misses.

Prefer [one-shot import](/docs/import-export/job-import) for a one-time migration file, and [backfill](/docs/backfill/configure-backfill) when you want Cavuno to scrape company careers pages instead of an XML feed.

## Related guides

- [Import jobs from CSV or XML](/docs/import-export/job-import): One-shot uploads
- [Configure backfill](/docs/backfill/configure-backfill): Careers-page aggregation
- [Job moderation](/docs/managing-jobs/job-moderation): Imports publish without the approval queue
- [Apply link tracking](/docs/board-setup/apply-link-tracking): Turn Off per feed when a destination rejects unexpected apply URL query params
