Personio Vacancy Watch | Jobs Scraper & Change Monitor

Watch known Personio boards for new, changed and removed jobs. This Personio jobs scraper reads public XML feeds and saves job rows or change events. It keeps the last complete board state when a check fails, so a failed check does not report every old job as removed.

Start with one board and snapshot mode. No Personio login is needed. Apify access is needed to run the Actor. A valid empty board or a delta with no changes can return zero rows. Platform usage can still have a cost; see Pricing and cost examples.

Table of contents

• What this Actor does
• What data you can get
• Quick start
• Full Input reference
• Input examples
• Output example
• Full Output field reference
• Pricing and cost examples
• Use cases
• Scheduling and monitoring
• API, MCP and integrations
• Limits and expected partial results
• Troubleshooting
• FAQ
• Related Actors
• Support

What this Actor does

Add known board names or exact HTTPS URLs on *.jobs.personio.de or *.jobs.personio.com. The Actor reads /xml?language=.... It does not search for companies, read applicant data, or open each job detail page.

Mode or event	What you receive
snapshot	Current accepted jobs, each marked baseline. Every complete snapshot also updates saved state.
First delta for a board and language	Current jobs marked baseline. There is no earlier complete state to compare.
Later delta: new	A job ID that was absent from the last complete snapshot.
Later delta: changed	A saved job whose normalized core content changed. changedFields lists the field names.
Later delta: removed	A saved job ID absent from a new complete feed. A removal is an observation, not proof that a vacancy was filled.
Later delta: unchanged	Same content; emitted only with includeUnchanged=true.

The monitor compares tenant|language inside the selected stateKey. Use different keys for independent monitors. The same tenant on .de and .com has the same state partition; use a separate stateKey if you need to compare these hosts independently.

What data you can get

Get job IDs, titles, office lists, department, employment type, seniority, schedule, creation date, descriptions, source salary values and public job URLs. Missing source values stay null or empty arrays. A language request can return no description if the employer has not supplied one in that language.

Existing consumers can read company, url, location, allLocations, postedAt, commitment and _scrapedAt. Their source meaning is listed below. Check these types and meanings before connecting an existing application. Full compatibility with every application is not promised.

team, country, workplaceType and updatedAt are null. The Actor does not infer a country or remote status from a title or office. Salary is returned only from published source salary content; it is not estimated from a role name.

Quick start

1. Open the Actor you have access to in Apify Console, then open Input.
2. Switch to JSON input and paste the minimal example below. Use a tenant from the company public Personio careers board, such as teamative. A job title or search word such as mediabuyer is not a verified tenant.
3. Keep language as en and choose snapshot for this first check.
4. Click Start. Wait for the run to finish.
5. Open Output or the run's default Dataset. Check tenant, jobId, title and eventType.
6. Open the default Key-value store and read RUN_SUMMARY. Check each board's status and complete.

If every board has a technical error, the run is Failed. Read RUN_SUMMARY in Storage; no new complete snapshot or removal events are saved. A real HTTP 404 is not_found and can finish without an exception. A redirect to the Personio home page is an error, not proof of an empty board.

In the captured example, teamative returned 18 jobs. Live jobs can change; 18 is an example count, not a required future result. For a repeated monitor, use delta and keep a stable stateKey. Start runs on the same key one at a time.

Full Input reference

The JSON object has 9 supported fields. Only boards is required. The fresh Input form offers teamative in snapshot mode, a 100-job cap and a separate PERSONIO_UI_DEMO_TEAMATIVE_EN state key. Repeated sample runs return current jobs without moving your normal monitor baseline. API calls still require boards; omitted optional fields use the runtime defaults below, including delta mode.

Field	Input label	Type	Required	Runtime default	UI prefill	Allowed values	Effect and recommendation
boards	Personio board names or URLs	array	Yes	No default; required. Start with one known board.	["teamative"]	1–50 non-empty strings	Reads each board in order. Use a slug such as teamative, or its exact HTTPS Personio board URL. Slugs use .de.
language	Job text language	string	No	en	None	de, en, es, fr, it, nl, pl, pt	Requests this language from the XML feed. Start with en; try the source language if text is empty. It does not translate text.
mode	Snapshot or changes	string	No	delta	snapshot	snapshot, delta	Snapshot returns current jobs as baseline rows. Delta compares saved state. Use snapshot for one-off exports and delta for repeated checks.
maxJobsPerBoard	Jobs allowed per board	integer	No	5000	100	integer 1–5000	If the board has more jobs, returns only this many with truncated status. It keeps old state and emits no removals. Choose a limit above your board size.
maxBodyBytes	XML size limit (bytes)	integer	No	10485760	None	integer 16384–10485760	Rejects an XML response above this size. Keeps old state. A 1 MiB limit (1048576) worked for the example boards; raise it for larger descriptions.
boardTimeoutMs	Time limit per board (ms)	integer	No	30000	None	integer 1000–120000	Limits the whole board fetch, including retries and waits. Keep 30000 first; a lower value can stop a slow board before all attempts.
maxAttempts	Request attempts per board	integer	No	3	None	integer 1–5	Includes the first request. Retries selected HTTP or transport failures within the board time limit. Use 1 to avoid retries or keep 3 for normal checks.
stateKey	Monitor state key	string	No	PERSONIO_MONITOR_STATE	PERSONIO_UI_DEMO_TEAMATIVE_EN	1–128 characters; pattern below	Selects a record in the named state store. Keep it stable for one monitor. Use a new key for an independent baseline.
includeUnchanged	Show unchanged jobs	boolean	No	false	None	boolean true or false	Adds unchanged rows in delta mode. Keep false for a change feed. It has no effect in snapshot mode.

stateKey accepts ^[A-Za-z0-9][A-Za-z0-9._:-]{0,127}$. INPUT, OUTPUT, RUN_SUMMARY and STATE_SNAPSHOT are reserved. Null is not a default. An empty boards array, blank strings, duplicate canonical boards or unknown fields cause an input error.

Board URLs accept only an empty path, / or /xml. A single language query is allowed and must match the input language. Job detail URLs, ports, fragments, credentials, custom domains and redirects are not accepted. A slug must be one plain ASCII tenant label. Language codes are normalized to lowercase.

Settings that work together

Settings	Result
language + URL query	A conflicting URL language is rejected. Language is part of saved board identity.
mode + includeUnchanged	includeUnchanged only adds rows in delta mode. Snapshot rows are baseline.
maxJobsPerBoard + state	A truncated feed can emit capped current events, but cannot replace saved state or emit removals.
maxAttempts + boardTimeoutMs	The board deadline includes retries and waits. More attempts cannot extend the deadline.
stateKey + board list	Boards not requested keep their old snapshots. Removing a board from input does not emit its jobs as removed.
snapshot + stateKey	A complete snapshot replaces that board's saved state. Use a separate key for one-off exports if you also run a delta monitor.

Filters and reset

The Actor filters board hosts, XML shape and safe source URLs for valid input. It has no title, department, location, remote, date or salary search filters. Use the exported fields for your own filtering after the run. The job cap is a safety limit, not a search filter.

To start a new baseline, stop overlapping runs and choose a new stateKey, such as WEEKLY_PERSONIO_EN_V2. The old record stays intact. There is no reset flag and no state deletion in this workflow. To return to a monitor, restore its original key; it compares with that key's last complete snapshot.

Input examples

Minimal snapshot

{
  "boards": [
    "teamative"
  ],
  "language": "en",
  "mode": "snapshot"
}

This requests one current English board. In the captured source scope, it yields baseline job rows. The example uses a safe state key; the captured input and full samples are retained in the repository.

Repeated three-board monitor

{
  "boards": [
    "teamative",
    "personio",
    "sipgate"
  ],
  "language": "en",
  "mode": "delta",
  "maxJobsPerBoard": 100,
  "maxBodyBytes": 1048576,
  "boardTimeoutMs": 30000,
  "maxAttempts": 2,
  "stateKey": "WEEKLY_PERSONIO_EN",
  "includeUnchanged": false
}

Use this input unchanged for later checks. The first complete check on this key is a baseline. The next unchanged check can return zero rows. Three real boards returned 22 jobs in the checked snapshot and delta baseline, followed by zero repeat events. A future board may have more jobs than 100; check truncated status and raise the cap when needed.

Output example

This full job JSON is from a real teamative snapshot. The timestamp is the capture time, not a promise of current freshness. English description text was absent; the salary and offices came from the source.

{
  "_type": "job",
  "_schemaVersion": "1.0",
  "_source": "personio",
  "tenant": "teamative",
  "jobId": "2681315",
  "title": "Junior SAP Funkčný Konsultant – SAP CE / SAP IS-U",
  "publicUrl": "https://teamative.jobs.personio.de/job/2681315",
  "applyUrl": null,
  "subcompany": "teamative Slovakia s.r.o.",
  "office": "SK - Bratislava",
  "additionalOffices": [
    "SK - Košice"
  ],
  "department": "Delivery",
  "recruitingCategory": "SK - application process IT/SAP",
  "employmentType": "permanent",
  "seniority": "entry-level",
  "schedule": "full-time",
  "yearsOfExperience": "1-2",
  "keywords": [
    "SAP IS-U,SAP CE,ISU,CE,Junior"
  ],
  "occupation": "enterprise_software_implementation_consulting",
  "occupationCategory": "it_software",
  "createdAt": "2026-06-22T11:07:27+00:00",
  "descriptionSections": [],
  "descriptionText": null,
  "salaryInformation": {
    "minimum": 2000,
    "maximum": null,
    "currency": "EUR",
    "period": "monthly"
  },
  "company": "teamative",
  "url": "https://teamative.jobs.personio.de/job/2681315",
  "location": "SK - Bratislava",
  "allLocations": [
    "SK - Bratislava",
    "SK - Košice"
  ],
  "postedAt": "2026-06-22T11:07:27.000Z",
  "commitment": "Full-time",
  "team": null,
  "country": null,
  "workplaceType": null,
  "updatedAt": null,
  "observedAt": "2026-10-01T11:43:33.910Z",
  "contentHash": "6b6d47abef37ae0392319d40644adbbefd20f280ac210875376111463339dbfe",
  "_scrapedAt": "2026-10-01T11:43:33.910Z",
  "eventType": "baseline",
  "changedFields": []
}

JSON output keeps arrays and objects. For export:

1. Open the finished run's default Dataset.
2. Choose Export and JSON, CSV or Excel (XLSX).
3. Keep the identity, title and event fields when selecting columns.
4. Check Unicode titles and office names in the downloaded file.

JSON, CSV and XLSX exports were checked on 18 real rows. Use JSON for full nested sections, salary and arrays. CSV/Excel flatten nested values; select columns for your consumer. A table view can hide columns without removing them from raw JSON. See Dataset export documentation.

Full Output field reference

There are 41 Dataset fields: 39 are always present, with nullable values where shown; previousObservedAt and removedAt are event-specific. Source fields and aliases describe the same job. Monitor fields describe this check.

Source job fields

Field	Type	Empty or absent?	Example	Meaning and provenance
_type	string	No	"job"	Fixed value job. Removed records are still job records.
_schemaVersion	string	No	"1.0"	Record contract version 1.0; not the upstream XML version.
_source	string	No	"personio"	Fixed source marker personio.
tenant	string	No	"teamative"	Validated board slug. Use tenant + jobId as the job identity.
jobId	string	No	"2681315"	Numeric XML position ID, stored as a string. Unique within a board.
title	string	No	Junior SAP Funkčný Konsultant – SAP CE / SAP IS-U	XML job name, trimmed. XML entities are decoded once; CDATA text stays literal.
publicUrl	string	No	"https://teamative.jobs.personio.de/job/2681315"	Built from the validated source host and jobId.
applyUrl	string / null	Yes: null	null	Source value only when it is a safe same-host HTTPS /job/{jobId} URL, with no query or fragment. No invented apply link.
subcompany	string / null	Yes: null	"teamative Slovakia s.r.o."	XML subcompany value. This may differ from the tenant slug.
office	string / null	Yes: null	"SK - Bratislava"	First office value from XML.
additionalOffices	array	Yes: []	["SK - Košice"]	Remaining office values and explicit additional offices. Empty array when absent.
department	string / null	Yes: null	"Delivery"	XML department. No department filter is applied.
recruitingCategory	string / null	Yes: null	"SK - application process IT/SAP"	XML recruiting category, kept separate from team.
employmentType	string / null	Yes: null	"permanent"	XML employment type; no fixed job-type vocabulary is added.
seniority	string / null	Yes: null	"entry-level"	XML seniority value; no seniority guess from the title.
schedule	string / null	Yes: null	"full-time"	XML schedule value, such as full-time.
yearsOfExperience	string / null	Yes: null	"1-2"	Source text, not a calculated number.
keywords	array	Yes: []	["SAP IS-U,SAP CE,ISU,CE,Junior"]	Source keyword values; empty array when absent.
occupation	string / null	Yes: null	"enterprise_software_implementation_consulting"	XML occupation value.
occupationCategory	string / null	Yes: null	"it_software"	XML occupation category value.
createdAt	string / null	Yes: null	"2026-06-22T11:07:27+00:00"	Original source creation date string. It is not a last-update time.
descriptionSections	array	Yes: []	[]	Source sections with title and content. Content can contain HTML; treat it as untrusted text.
descriptionText	string / null	Yes: null	null	Text built from section headings and content with HTML tags removed. HTML character references are decoded once. Original HTML stays in descriptionSections. Null when no useful text is available.
salaryInformation	object / null	Yes: null	{"minimum": 2000, "maximum": null, "currency": "EUR", "period": "monthly"}	Published salary values, including supported currencyCode/type names. Numbers may be read from source salary text. No salary prediction or currency conversion.

Existing consumer aliases

Field	Type	Empty or absent?	Example	Meaning and provenance
company	string	No	"teamative"	tenant alias. A board identifier, not a verified legal company name.
url	string	No	"https://teamative.jobs.personio.de/job/2681315"	publicUrl alias.
location	string / null	Yes: null	"SK - Bratislava"	office alias.
allLocations	array / null	Yes: null	["SK - Bratislava", "SK - Košice"]	Unique office values, main office first. Null if no offices are supplied.
postedAt	string / null	Yes: null	"2026-06-22T11:07:27.000Z"	An ISO date-time createdAt converted to UTC only when its calendar date, time and explicit timezone are valid. Null for a missing, invalid or timezone-free value. The original createdAt stays unchanged.
commitment	string / null	Yes: null	"Full-time"	schedule with the first letter in uppercase. Null when absent.
_scrapedAt	string	No	"2026-10-01T11:43:33.910Z"	Time of this check; on removed rows, this differs from the last-seen observedAt.
team	null	Yes: null	null	Always null: there is no dedicated mapped source field.
country	null	Yes: null	null	Always null: no country is inferred from office names.
workplaceType	null	Yes: null	null	Always null: no remote or hybrid status is inferred.
updatedAt	null	Yes: null	null	Always null: no reliable source update timestamp is mapped.

Monitor extensions

Field	Type	Empty or absent?	Example	Meaning and provenance
eventType	string	No	baseline / new / changed / removed / unchanged	baseline, new, changed, removed or unchanged; see mode rules.
observedAt	string	No	"2026-10-01T11:43:33.910Z"	Run-start UTC time for a current job. For removed rows, keeps the last time that job was seen.
changedFields	array	Yes: []	[]	Names of changed core fields, only for changed events. Empty array for other events. Alias names are not repeated.
contentHash	string	No	6b6d47abef37… (64 hex characters)	SHA-256 of 24 normalized core fields. Stable for unchanged content. Observation times and aliases are excluded.
previousObservedAt	string	Absent on other events	ISO UTC; changed or removed only	Previous saved observation for changed or removed events. Absent for other event types.
removedAt	string	Absent on other events	ISO UTC; removed only	UTC time when a complete feed no longer contains the job. Present only for removed events. It does not prove a role was filled.

Nested job values

Nested field	Type	Meaning / empty value / example
descriptionSections[].title	string / null	Section name, or null. Example: Your profile.
descriptionSections[].content	string	Source section body; may contain HTML. Example: Public job text . Never execute it.
salaryInformation.minimum	number / null	Source minimum, e.g. 2000.
salaryInformation.maximum	number / null	Source maximum; null in the selected live row.
salaryInformation.currency	string / null	Source currency/currencyCode or recognizable source currency text, e.g. EUR.
salaryInformation.period	string / null	Source period/type or recognizable source period text, e.g. monthly.

XML & becomes & and & becomes &. CDATA is literal at the XML layer. Description text is then cleaned as HTML text, so its HTML entities follow text-cleaning rules. DOCTYPE/ENTITY declarations and malformed feeds are refused. These cases do not become valid empty boards.

The content hash includes 24 original core fields, with normalized whitespace and stable object key order. Arrays keep their order. Observation timestamps and consumer aliases do not trigger changed events. A section or office change can change the hash. changedFields lists original core names, such as office, not alias names such as location.

Where each result is saved

Surface	What it contains
Default Dataset	All emitted job events from this run. Error details are in summaries, not extra job/error rows.
Default KVS INPUT	The run input.
Default KVS OUTPUT	Compact result, counts, board results and first up to 10 events.
Default KVS RUN_SUMMARY	Full board outcomes, attempts, resource counters and billing status.
Default KVS STATE_SNAPSHOT	Captured monitor state for this run, including saved snapshots from earlier checks.
Named KVS personio-jobs-monitor-state	Durable state under the chosen stateKey. Separate runs use this store for comparisons.
Output link jobs	Default Dataset API items.
Output link runOutput / runSummary / state	OUTPUT / RUN_SUMMARY / STATE_SNAPSHOT in the run's default KVS.

An input error writes OUTPUT and RUN_SUMMARY and fails the process. A PPE persistence or billing failure writes BILLING_ERROR and fails the process; saved records stay available. If data writing stopped, OUTPUT, RUN_SUMMARY or STATE_SNAPSHOT may be absent. Check the Dataset, receipt and platform status too. Other runtime errors write error surfaces before the process fails. STATE_SNAPSHOT may be absent. An input error summary uses fallback mode/language/key values, not validated input. A partial board-check loop can finish with process success; inspect each board before using its rows. When every board has a technical error, or the run is cancelled, the process fails after saving diagnostics. An all-404 result can finish successfully with explicit not_found statuses, but it is not a complete empty XML snapshot. Always inspect the summary.

The sanitized KVS examples below use a replaced run ID and state key. The reduced state sample contains one job from the 18-job snapshot; it is not the full snapshot. Its job values and shape are unchanged.

OUTPUT

Field	Type / presence	Meaning / example
schemaVersion	string	Fixed personio-monitor-output/v1.
runId	string	Apify run ID; EXAMPLE_RUN_ID in the sanitized sample.
terminalOutcome	string	success, partial, error or cancelled. Read this before using rows.
mode	string	snapshot or delta for this run.
language	string	Requested language, e.g. en.
observedAt	string	Run-start UTC timestamp.
datasetItemCount	integer	Number of emitted Dataset rows, e.g. 18.
eventCounts	object	Emitted counts by event type; see EventCounts below.
boards	array	One BoardSummary for each requested board.
dataset	object	Default Dataset selector and emitted item count.
dataset.key	string	Fixed default; not an Apify Dataset ID.
dataset.itemCount	integer	Same emitted row count as datasetItemCount.
sample	array	First up to 10 compact event rows; empty when no rows are emitted.
sample[].eventType	string	Event classification.
sample[].tenant	string	Board slug.
sample[].jobId	string	Source job ID.
sample[].title	string	Source job title.
sample[].contentHash	string	SHA-256 core hash.

RUN_SUMMARY

Field	Type / presence	Meaning / example
schemaVersion	string	Fixed personio-monitor-run-summary/v1.
runId	string	Apify run ID.
mode	string	Run mode.
language	string	Requested language.
stateKey	string	Selected record in the named state store.
startedAt	string	Run-start UTC timestamp.
finishedAt	string	Monitor-finish UTC timestamp.
durationMs	number	Monitor elapsed time, e.g. 734; not the full platform run duration.
terminalOutcome	string	success, partial, error or cancelled.
boardCount	integer	Number of board summaries.
successfulBoardCount	integer	Boards with status success, including valid empty feeds.
failedBoardCount	integer	Boards with other errors; truncated boards are counted separately.
truncatedBoardCount	integer	Boards with too many positions for maxJobsPerBoard.
datasetItemCount	integer	Number of emitted events.
eventCounts	object	Counts of emitted events, not all unchanged jobs.
boards	array	BoardSummary objects below.
accounting	object	Resource counters and billing status below.
redaction	object	Logging flags below; not a promise that source job text has no contacts.
errorClass	string; optional	Top-level runtime/input/storage error code; absent on normal monitor completion.
errorMessage	string; optional	Safe top-level diagnostic; absent on normal completion.

EventCounts

Field	Type / presence	Meaning / example
baseline	integer	Number of emitted baseline rows; zero when none are emitted.
new	integer	Number of emitted new rows; zero when none are emitted.
changed	integer	Number of emitted changed rows; zero when none are emitted.
removed	integer	Number of emitted removed rows; zero when none are emitted.
unchanged	integer	Number of emitted unchanged rows; zero when none are emitted.

BoardSummary (OUTPUT.boards[] and RUN_SUMMARY.boards[])

Field	Type / presence	Meaning / example
key	string	tenant|language, e.g. teamative|en. Host suffix is not a separate state partition.
tenant	string	Canonical tenant slug.
language	string	Requested language.
sourceUrl	string	Canonical HTTPS XML request URL.
status	string	success, truncated, not_found, access_block, transport_error, schema_drift, timeout, cancelled, security_error or body_cap.
domainStatus	string / null	empty or non_empty after XML parsing; null on a failed fetch.
complete	boolean	Whether the entire feed was accepted. Only true feeds replace saved board state.
truncated	boolean	True when job count exceeds the cap.
cap	integer	Input maxJobsPerBoard.
jobCount	integer	Accepted positions, capped on a truncated feed; not the discarded total.
emittedItemCount	integer	Rows emitted from this board; may include removals in delta.
attempts	integer	Attempt-log length, including the first request.
durationMs	number	Elapsed time for this board.
errorClass	string / null	Failure code, or null on parsed success/truncation.
errorMessage	string / null	Safe diagnostic, or null on parsed success/truncation.
attemptLog	array	AttemptRecord objects below; [] if cancelled before fetching.

AttemptRecord (boards[].attemptLog[])

Field	Type / presence	Meaning / example
attempt	integer	1-based request attempt number.
status	integer / null	HTTP status, e.g. 200; null if no status was captured.
durationMs	number	Duration of this attempt.
result	string	response, retry, failure or cancelled.
errorClass	string; optional	Diagnostic code on an error/retry; absent on a normal response.
retryAfterMs	number / null; optional	Server Retry-After parsed and bounded to 10000 ms; absent on other paths, null if unparseable.

Accounting (RUN_SUMMARY.accounting)

Field	Type / presence	Meaning / example
currency	string	Fixed USD.
cloudChargeUsd	number / null	Null in Cloud until platform read-back; local runs set 0 because no Cloud run is billed.
billingStatus	string	pending_platform_readback in Cloud; not_applicable locally. Read final run usage in Console for a platform cost.
externalSpendUsd	number	Fixed 0: the Actor does not call paid external services. This is not total platform billing.
localResourceAccounting	object	Counters below; not dollar amounts.
localResourceAccounting.networkRequests	integer	Number of recorded board attempts; retries included.
localResourceAccounting.responseBytes	integer	Bytes counted from returned responses. A response interrupted inside the transport may not be counted here.
localResourceAccounting.datasetItems	integer	Number of emitted Dataset rows.

Redaction (RUN_SUMMARY.redaction)

Field	Type / presence	Meaning / example
rawResponsesLogged	boolean	Fixed false: raw XML is not written to runtime logs.
credentialsLogged	boolean	Fixed false: source credentials are not accepted or logged.
contactDetailsLogged	boolean	Fixed false for routine log messages. Public job descriptions in Dataset/state can still contain contacts.

MonitorState (named stateKey and default STATE_SNAPSHOT)

Field	Type / presence	Meaning / example
schemaVersion	string	Fixed personio-monitor-state/v1.
updatedAt	string	Last monitor-finish UTC timestamp; may advance while failed board snapshots stay unchanged.
snapshots	object	Map by tenant|language; includes earlier boards not requested in this run.
snapshots.<tenant|language>	object	StoredBoardSnapshot below. Missing if this board has never completed successfully.

StoredBoardSnapshot

Field	Type / presence	Meaning / example
tenant	string	Board slug.
language	string	Requested language.
sourceUrl	string	Accepted XML feed URL.
observedAt	string	Time this complete board was last seen.
jobs	object	Map by jobId. Empty object for a complete empty feed.
jobs.<jobId>	object	All JobCore source fields and aliases above, plus observedAt and contentHash. No eventType, changedFields, _scrapedAt, previousObservedAt or removedAt.

OUTPUT sample

{
  "schemaVersion": "personio-monitor-output/v1",
  "runId": "EXAMPLE_RUN_ID",
  "terminalOutcome": "success",
  "mode": "snapshot",
  "language": "en",
  "observedAt": "2026-10-01T11:43:33.910Z",
  "datasetItemCount": 18,
  "eventCounts": {
    "baseline": 18,
    "new": 0,
    "changed": 0,
    "removed": 0,
    "unchanged": 0
  },
  "boards": [
    {
      "key": "teamative|en",
      "tenant": "teamative",
      "language": "en",
      "sourceUrl": "https://teamative.jobs.personio.de/xml?language=en",
      "status": "success",
      "domainStatus": "non_empty",
      "complete": true,
      "truncated": false,
      "cap": 100,
      "jobCount": 18,
      "emittedItemCount": 18,
      "attempts": 1,
      "durationMs": 586,
      "errorClass": null,
      "errorMessage": null,
      "attemptLog": [
        {
          "attempt": 1,
          "status": 200,
          "durationMs": 566,
          "result": "response"
        }
      ]
    }
  ],
  "dataset": {
    "key": "default",
    "itemCount": 18
  },
  "sample": [
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "2819528",
      "title": "Consultant Informationssicherheit / Notfallmanagement (BCM) (m/w/d)",
      "contentHash": "98e456861e61a43c4900488870986d11809756fdf66c3ad1254f691a98aaa53e"
    },
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "2623782",
      "title": "Initiativbewerbung",
      "contentHash": "8e7d003519b5838ca1f61b6c1b2dde278fb7870415f5321248372431dcb3b069"
    },
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "2681315",
      "title": "Junior SAP Funkčný Konsultant – SAP CE / SAP IS-U",
      "contentHash": "6b6d47abef37ae0392319d40644adbbefd20f280ac210875376111463339dbfe"
    },
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "2805114",
      "title": "Lead Consultant SAP Architecture",
      "contentHash": "80861f66a6e54c531a6479c639d12f2242209d8e12905aefb59617caf2067aab"
    },
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "2617009",
      "title": "Sales Manager – ServiceNow",
      "contentHash": "68a21a83ee5658480905f028382ad56acf5cfcc31ce0e10f01bc46689ccaeec8"
    },
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "2231198",
      "title": "SAP CS Consultant with German",
      "contentHash": "c0ed5e0aa47157018b4b41e45a448afe6bc3836762fec57496821da2ef0520be"
    },
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "2232665",
      "title": "SAP EWM Consultant",
      "contentHash": "457f43f34971f3386bef18eb82188351dd0cf942ef78d781876142b96476addc"
    },
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "2670413",
      "title": "SAP IS-U Consultant with German (m/f/d)",
      "contentHash": "f3ae88c10b769487618d9a34566376bee59a840f162d7d29595439017022eef7"
    },
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "1896335",
      "title": "SAP IS-U Konzultant pro Českou republiku (EU-remote)",
      "contentHash": "6288fabc6ff238335bafca8ee8ddc6271450e501484b28ced77e1e12031dae28"
    },
    {
      "eventType": "baseline",
      "tenant": "teamative",
      "jobId": "2553655",
      "title": "Senior Consultant ServiceNow ITOM / CMDB",
      "contentHash": "8254686597814b510ffe3b6fb0b255306c2cf6df9cda39d37f6b0b6eec315989"
    }
  ]
}

RUN_SUMMARY sample

{
  "schemaVersion": "personio-monitor-run-summary/v1",
  "runId": "EXAMPLE_RUN_ID",
  "mode": "snapshot",
  "language": "en",
  "stateKey": "PERSONIO_EXAMPLE_SNAPSHOT",
  "startedAt": "2026-10-01T11:43:33.910Z",
  "finishedAt": "2026-10-01T11:43:34.644Z",
  "durationMs": 734,
  "terminalOutcome": "success",
  "boardCount": 1,
  "successfulBoardCount": 1,
  "failedBoardCount": 0,
  "truncatedBoardCount": 0,
  "datasetItemCount": 18,
  "eventCounts": {
    "baseline": 18,
    "new": 0,
    "changed": 0,
    "removed": 0,
    "unchanged": 0
  },
  "boards": [
    {
      "key": "teamative|en",
      "tenant": "teamative",
      "language": "en",
      "sourceUrl": "https://teamative.jobs.personio.de/xml?language=en",
      "status": "success",
      "domainStatus": "non_empty",
      "complete": true,
      "truncated": false,
      "cap": 100,
      "jobCount": 18,
      "emittedItemCount": 18,
      "attempts": 1,
      "durationMs": 586,
      "errorClass": null,
      "errorMessage": null,
      "attemptLog": [
        {
          "attempt": 1,
          "status": 200,
          "durationMs": 566,
          "result": "response"
        }
      ]
    }
  ],
  "accounting": {
    "currency": "USD",
    "cloudChargeUsd": null,
    "billingStatus": "pending_platform_readback",
    "externalSpendUsd": 0,
    "localResourceAccounting": {
      "networkRequests": 1,
      "responseBytes": 44712,
      "datasetItems": 18
    }
  },
  "redaction": {
    "rawResponsesLogged": false,
    "credentialsLogged": false,
    "contactDetailsLogged": false
  }
}

Reduced STATE_SNAPSHOT sample

{
  "schemaVersion": "personio-monitor-state/v1",
  "updatedAt": "2026-10-01T11:43:34.644Z",
  "snapshots": {
    "teamative|en": {
      "tenant": "teamative",
      "language": "en",
      "sourceUrl": "https://teamative.jobs.personio.de/xml?language=en",
      "observedAt": "2026-10-01T11:43:33.910Z",
      "jobs": {
        "2681315": {
          "_type": "job",
          "_schemaVersion": "1.0",
          "_source": "personio",
          "tenant": "teamative",
          "jobId": "2681315",
          "title": "Junior SAP Funkčný Konsultant – SAP CE / SAP IS-U",
          "publicUrl": "https://teamative.jobs.personio.de/job/2681315",
          "applyUrl": null,
          "subcompany": "teamative Slovakia s.r.o.",
          "office": "SK - Bratislava",
          "additionalOffices": [
            "SK - Košice"
          ],
          "department": "Delivery",
          "recruitingCategory": "SK - application process IT/SAP",
          "employmentType": "permanent",
          "seniority": "entry-level",
          "schedule": "full-time",
          "yearsOfExperience": "1-2",
          "keywords": [
            "SAP IS-U,SAP CE,ISU,CE,Junior"
          ],
          "occupation": "enterprise_software_implementation_consulting",
          "occupationCategory": "it_software",
          "createdAt": "2026-06-22T11:07:27+00:00",
          "descriptionSections": [],
          "descriptionText": null,
          "salaryInformation": {
            "minimum": 2000,
            "maximum": null,
            "currency": "EUR",
            "period": "monthly"
          },
          "company": "teamative",
          "url": "https://teamative.jobs.personio.de/job/2681315",
          "location": "SK - Bratislava",
          "allLocations": [
            "SK - Bratislava",
            "SK - Košice"
          ],
          "postedAt": "2026-06-22T11:07:27.000Z",
          "commitment": "Full-time",
          "team": null,
          "country": null,
          "workplaceType": null,
          "updatedAt": null,
          "observedAt": "2026-10-01T11:43:33.910Z",
          "contentHash": "6b6d47abef37ae0392319d40644adbbefd20f280ac210875376111463339dbfe"
        }
      }
    }
  }
}

Optional PPE records

These records appear only when this run uses the supported pay-per-event model. They are in the default KVS, separate from your named monitor state. A receipt confirms platform event counts; it is not a customer invoice or proof of developer profit.

Record / field	Type / example	Meaning
BILLING_RECEIPT.schemaVersion	string / personio-monitor-billing/v1	Receipt format.
BILLING_RECEIPT.runId	string / run ID	Exact run; a different run cannot reuse it.
BILLING_RECEIPT.digest	string / 64 hex characters	SHA-256 of the result and planned counts.
BILLING_RECEIPT.phase	saving / persisted / confirmed	Data write started, data saved, or all event counts confirmed.
BILLING_RECEIPT.counts.complete-board-check	integer / 1	Complete successful boards, including valid empty feeds.
BILLING_RECEIPT.counts.saved-current-job	integer / 18	Unique current-job events from complete boards. Removed events are free.
BILLING_RECEIPT.calls.complete-board-check	ready / in_flight / confirmed	Board charge call not sent, uncertain, or confirmed by platform counts.
BILLING_RECEIPT.calls.saved-current-job	ready / in_flight / confirmed	Current-job charge call state.
BILLING_ERROR.runId	string / run ID	Run that failed during persistence or billing.
BILLING_ERROR.errorClass	string / storage_error	Safe failure class.
BILLING_ERROR.errorMessage	string / safe message	Reason to inspect results and platform counts.

If a charge response is lost, the Actor reads platform counts. It does not send an uncertain charge again. If data saving was interrupted, it stops recovery to avoid duplicate rows. Preserve the run and contact support. Do not keep restarting it. A later independent run uses your saved monitor state and has its own billable checks.

Pricing and cost examples

Check the Pricing tab and your run limit before Start. This Actor uses pay per event. The Pricing tab shows whether platform usage is included or charged separately.

Event	Price in USD	What you pay for
Complete board check	0.018	One complete successful board, including a valid empty feed.
Saved current job	0.00045	One unique baseline, new, changed or requested unchanged job from a complete board, saved to the Dataset.

Removed jobs, failed/truncated boards and their job rows have no event fee. There is no start fee or extra default Dataset event fee. A check with no changes still has the board fee. For one complete board and18 current jobs, events costUSD0.0261. For three complete boards and22 current jobs, events costUSD0.0639. For three complete boards with no emitted current jobs, events costUSD0.054. These examples describe event charges; your actual counts may differ.

When the Pricing tab shows pay per event + usage, Apify charges platform usage separately, including for failed or partial runs. When usage is included, it is covered by the event price. A zero event fee does not promise zero platform resource use. Check both event fees and usage before scheduling.

The Actor saves results before charging. If the full result exceeds your run charge limit, the run fails without saving or charging that result. Source requests may already have used resources. If a billing call fails after saving, inspect BILLING_RECEIPT and BILLING_ERROR; do not repeatedly restart an uncertain charge.

The observations below are historical platform usage from private technical runs; they are not customer debit evidence for this PPE version.

Observed scope	Emitted rows	Platform usage USD
One teamative snapshot, first checked run	18	0.00045294
Three boards, snapshot	22	0.00044319
Three boards, delta baseline	22	0.00044465
Same three boards, unchanged delta	0	0.00031457

These runs used 256 MB and final platform usage read-back. They exclude build costs and future extra reads/exports. Billing meters can update after a run ends; RUN_SUMMARY.cloudChargeUsd stays null in Cloud. A zero-row delta still requests boards and reads/writes state. Fewer output rows do not mean a free check.

Use a platform run charge limit, a sensible timeout, and one test board before scheduling. maxJobsPerBoard limits accepted positions per board; it is not a dollar budget or a cap on removed events. No Cloud cost is claimed for 1000 or 5000 jobs from these small runs.

Use cases

Buyer task	Input	Useful fields	Next step
Keep a current vacancy list	snapshot on known boards	tenant, jobId, title, publicUrl	Export JSON or CSV; replace your current list only from complete boards.
Track added or edited jobs	delta with a stable stateKey	eventType, changedFields, contentHash	Upsert by tenant + jobId in your database.
Mark a job as no longer listed	delta with complete source	removedAt, previousObservedAt, jobId	Archive the saved role; keep its last-known content.
Check locations and published pay	either mode	allLocations, salaryInformation	Filter these values in your own worksheet or application.

Scheduling and monitoring

These are setup steps for your account, not a schedule created by this Actor.

1. Run and check the intended delta input once.
2. In Apify Console, open Schedules and Create new.
3. Set the timezone and frequency, for example once per day.
4. Select this Actor or a saved task with the same boards, language and stateKey.
5. Select a tested build and set timeout, memory and charge limits.
6. Leave enough time between runs for the full board list to finish. Do not overlap runs on the same key.
7. After each run, read RUN_SUMMARY. Use its board complete/status fields before treating missing rows as domain events.

See Apify scheduling. Scheduled events can be delayed. This Actor has no state lock or automatic alert delivery; send notifications from your own workflow after checking eventType and terminalOutcome.

API, MCP and integrations

API

The API example uses this Actor's exact ID. Set PERSONIO_BUILD to the exact build number you have checked in Console. It does not use a moving build tag. Your account must have access to the Actor. Start creates a run and can incur a charge. Keep your token in APIFY_TOKEN in the environment. Do not put it in a URL, source file, screenshot or shared input.

# APIFY_TOKEN must already be set securely in your environment.
# Set PERSONIO_BUILD to your checked build number (for example, 0.1.4 for the earlier runtime).
: "${PERSONIO_BUILD:?Set the exact checked build number first}"
curl --fail-with-body --request POST \
  "https://api.apify.com/v2/acts/vJ9HvIdGAVEGWQbxF/runs?build=${PERSONIO_BUILD}&memory=256&timeout=120&maxTotalChargeUsd=0.03" \
  --header "Authorization: Bearer ${APIFY_TOKEN:?Set APIFY_TOKEN securely first}" \
  --header 'Content-Type: application/json' \
  --data '{"boards":["teamative"],"language":"en","mode":"snapshot","stateKey":"API_PERSONIO_EN"}'

Read data.id from the response. Poll GET /v2/actor-runs/{RUN_ID} until it is terminal. Read data.defaultDatasetId and data.defaultKeyValueStoreId. Download GET /v2/datasets/{DATASET_ID}/items?format=json and GET /v2/key-value-stores/{KVS_ID}/records/RUN_SUMMARY, with the same Authorization header. For tabular exports use format=csv or format=xlsx. Check both platform status and the summary. See Run Actor API.

MCP and downstream tools

Apify provides an MCP server for supported Actors and storage tools. This Actor has no custom MCP server, and its execution through MCP has not been verified. Use the API above for the documented path. An existing authorized MCP storage tool may read a Dataset that your account can access; check that tool's permissions first.

Your own Node.js/Python app can read the Dataset API. You can also import CSV/XLSX into a spreadsheet. For an upsert, use tenant + jobId; use eventType for the action and contentHash for content checks. This Actor does not create a database, CRM record, webhook or notification service.

Limits and expected partial results

Input safety caps are 50 boards, 5000 positions per board and 10 MiB per response. Real Cloud checks covered up to 22 current jobs across 3 boards at 256 MB. A local synthetic test accepted 5000 positions on one board. The local test is not a Cloud capacity or throughput guarantee. Start small and inspect the results for your source.

Situation	Dataset	Saved board state	How to read it
Complete feed with no jobs	No current jobs; delta may emit removals	Replaced with an empty job map	status=success, domainStatus=empty, complete=true.
Unchanged delta	Empty unless includeUnchanged=true	Updated complete snapshot	success and zero emitted events are normal.
More jobs than the cap	Capped current events; no removals	Old snapshot kept	status=truncated, complete=false, terminalOutcome=partial.
404, block, bad XML, body cap or timeout	No events from that failed board	Old snapshot kept	Error status, complete=false; not an empty domain result.
Some boards fail	Successful/capped board events stay available	Only complete boards replace their state	terminalOutcome=partial; inspect boards individually.
All boards fail	No job events	Old board snapshots kept	terminalOutcome=error, even if platform process completed.

Snapshots measure the source only at check time. Language availability, published fields, source changes and rate limits can affect output. Retry delays are bounded; only selected failures retry. Redirects and private/reserved network addresses are refused. No proxy or Personio credential input is supported.

The Actor reads public job feeds. It does not read applications, candidate profiles, private HR records or login-only data. Public descriptions may contain names, email addresses or phone numbers. They are not automatically redacted from Dataset/state. Review them before sharing. Follow source terms and your data-use duties. This is an independent product, not affiliated with or endorsed by Personio.

Troubleshooting

Symptom	Likely reason	Check	Safe next action
Input error before requests	Bad board shape, duplicate board, reserved state key or unknown field	Input and top-level errorClass	Use the exact fields and valid board root URL.
Failed run with redirect_blocked	The tenant redirects away from its XML feed	RUN_SUMMARY and the company careers board	Use the actual public company tenant. Redirects stay blocked. Previous state is kept.
Empty delta Dataset	No changes, a valid empty board, or failed checks	terminalOutcome, status, domainStatus, complete	For no changes do nothing; for failure fix the input/source and retry.
descriptionText is null	Employer has no text in the requested language	descriptionSections and public source	Try the employer's language with a separate monitor key if desired.
Posted time is null but createdAt is filled	No explicit timezone or invalid source date	Original createdAt string	Keep the source value; do not invent a timezone.
Truncated status	Feed exceeds maxJobsPerBoard	cap and jobCount	Raise the cap within its range; previous complete state stays intact.
body_cap / timeout	Response is too large or slow	maxBodyBytes, boardTimeoutMs and attemptLog	Raise the relevant bound within the input range, then test one board.
not_found / access_block / schema_drift	Wrong board, denied access or changed XML	sourceUrl and safe errorMessage	Check the public board; report a reproducible source change. Do not treat it as all jobs removed.
Repeated new/changed events	Truncated snapshots, a changed key, or overlapping runs	complete, stateKey and run times	Keep one key, complete feeds and sequential runs.
State cannot be read	Unsupported or damaged saved state	Top-level error and named KVS record	Preserve the old record. Use a new key for a fresh baseline, then report the issue.

FAQ

Do I need a Personio account?

No. Only known public XML boards are read. You need Apify access to run the Actor. Private job data is outside this workflow.

Can I use a custom careers domain or a job URL?

No. Use a plain board slug or the supported Personio board root/XML URL. A slug points to .de; use a full URL for .com.

Does removed mean the company filled the job?

No. It means the ID disappeared from a complete accepted feed. The employer may have hidden, moved or removed the listing for another reason.

Why is the first delta a baseline?

There is no prior complete board snapshot under this key and language. Later delta runs compare with the saved state. Changing the key starts a new comparison.

Does snapshot mode affect my monitor?

Yes. A complete snapshot updates the selected state record. Use a separate key for an export that should not move your delta baseline.

After an update that corrects derived fields, use snapshot mode to refresh all exported rows. Delta reports source content changes; it does not refresh every unchanged row. Use a separate key if you want to keep your existing monitor baseline.

Is createdAt the job's last change date?

No. It is the source creation value. updatedAt remains null. changed events compare content; they do not supply an employer update timestamp.

Are all fields always filled?

No. Null means the mapped value is absent or not safe to use. Empty arrays mean no values were supplied. Optional event timestamps can be absent. A job row with null source fields is different from a failed board fetch.

Are repeated runs free if no jobs change?

No free-run promise is made. Source checks and state work still use platform resources. Check the Pricing tab and run usage.

Related Actors

• Greenhouse Jobs Scraper & Monitor — use the owner's published Greenhouse tool when the source board is on Greenhouse.
• Lever Job Scraper — use the owner's published Lever tool for public Lever boards.

These are separate source tools. Check each tool's input and output before joining their data.

Support

Open the Actor's Issues tab for a question or reproducible bug. This is the support route listed by Cliqto Media. Include the run ID, board URL, language, mode and safe error text. Share only the fields needed to reproduce the issue. Never post tokens, cookies, private payloads or account billing details.