> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.beatapi.io/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.beatapi.io/_mcp/server.

# How do I find recent remote LinkedIn jobs, and why does pagination not advance by 50?

> Search recent remote LinkedIn jobs with linkedin.web_v2.search_jobs, send one value per filter, and page with start 0, 25, 50.

To find recent remote LinkedIn jobs, call `linkedin.web_v2.search_jobs` with `date_posted: "Past week"`, `onsite_remotes: ["Remote"]` and `sort_by: "Most recent"`. Pagination advances by 25, not 50: pass `start` as 0, 25, 50. Job search has no `pagination_token`; the step of 50 with a token belongs to personal and company post lists.

## How do I combine past week, remote and newest?

In `linkedin.web_v2.search_jobs`, set `date_posted` to `Past week`, `onsite_remotes` to `["Remote"]` and `sort_by` to `Most recent`. `geo_code` is an integer; `92000000` means worldwide, not a city name. Send the English values with the documented capitalization.

A complete first-page request for recent, remote, entry-level, full-time software roles is:

```json
{
  "action": "linkedin.web_v2.search_jobs",
  "params": {
    "keywords": "software engineer",
    "geo_code": 92000000,
    "date_posted": "Past week",
    "onsite_remotes": ["Remote"],
    "experience_levels": ["Entry level"],
    "job_types": ["Full-time"],
    "sort_by": "Most recent",
    "start": 0
  }
}
```

`Remote` is the job's work-mode filter. A worldwide search does not establish that you can work from any country, meet residency requirements or hold work authorization. Read the actual location and eligibility wording in each shortlisted job before deciding to apply. An entry-level label does not guarantee that the job description requires no experience.

## How are company, experience and job-type filters encoded?

The input schema accepts arrays, but arrays are not OR filters. For `company_ids`, `experience_levels`, `onsite_remotes`, `industries` and `job_types`, only the first value takes effect in the current service behavior. Send one intended value per field, such as `onsite_remotes: ["Remote"]`. `company_ids` takes integer IDs, not company names.

`title_ids` and `functions` are accepted by the input shape but do not filter results, so do not rely on `title_ids` or `functions` to narrow a search; put the text in `keywords` instead. `easy_apply` and `under_10_applicants` are booleans, not the string `"true"`. The gateway accepts `title_ids` and `functions` without an error; the two fields just have no effect.

## How do I search both Remote and Hybrid without losing half the filter?

Do not send `["Remote", "Hybrid"]` and expect both values to apply. Prepare two independent first-page requests with all other filters held constant:

```json
{ "action": "linkedin.web_v2.search_jobs", "params": { "keywords": "software engineer", "geo_code": 92000000, "date_posted": "Past week", "onsite_remotes": ["Remote"], "sort_by": "Most recent", "start": 0 } }
```

```json
{ "action": "linkedin.web_v2.search_jobs", "params": { "keywords": "software engineer", "geo_code": 92000000, "date_posted": "Past week", "onsite_remotes": ["Hybrid"], "sort_by": "Most recent", "start": 0 } }
```

Save the two bodies as `remote-request.json` and `hybrid-request.json`; each body is a separate call, not one batch request. The free Inspect check currently lists `$0.15` per search request, so the two first pages cost an estimated `$0.30` if both succeed. Confirm the current price and a two-call budget before sending. Further pages and job-detail calls cost extra.

Save each response with its own request ID, query time and source filter. When you later combine reviewed job records, deduplicate by the actual job identity or URL, and keep which search returned each job. Do not deduplicate by title, and do not treat a job found by both searches as proof that the role is both remote and hybrid; read the listing's work-location conditions before shortlisting the job. Page each search independently: `start: 0`, then `start: 25` if another page is approved. A two-page sample is not an exhaustive market scan.

## What are the first and second offsets?

Pass `start` as 0, 25, 50. Job search has no `pagination_token` parameter, so do not reuse the personal-post step of 50 plus a token. Keep the filters unchanged across pages, cap the page count, and stop on duplicate pages or business errors. Reaching your page cap ends your batch; the cap does not mean LinkedIn has no more matching jobs.

## What should I change when a narrow search returns no matches?

First tell apart a successful empty result, a failed response and an incomplete preview. Read the saved complete response before paying for a new query, and keep the original filters and observation time with the saved response.

If a remote search for the past week is empty, keep the remote filter and change only `Past week` to `Past month`. Start again at `start: 0`, and save the two filter sets and their results separately. If the new results are two or three weeks old, check each listing to confirm the job is still open.

`easy_apply: true` and `under_10_applicants: true` each add a restriction. Leaving `easy_apply` or `under_10_applicants` false adds no restriction; false does not mean “exclude Easy Apply” or “only jobs with at least ten applicants.” Every request with changed filters is a new paid query, unlike exporting data you already saved.

## How do I obtain required skills and the hiring team?

Pass an actual job URL to `linkedin.web_v2.get_job_detail`, and set `include_skills` or `include_hiring_team` to true when you need skills or the hiring team; both default to false. Each detail retrieval is a separate call, not part of the search. `linkedin.web_v2.search_jobs` and `linkedin.web_v2.get_job_detail` only read jobs; neither action applies to a job or contacts recruiters.

Sources: the gateway input contracts `data:linkedin.web_v2.search_jobs` and `data:linkedin.web_v2.get_job_detail`. The gateway returns `object`/`status`/`action`/`request_id`/`data` and does not project `data` into uniform job columns, so save the response as returned and do not invent a `jobs[]` path.

[Complete search-and-save workflow](/guides/linkedin-job-search)