Teamtailor Job Watch

Get public Teamtailor jobs from the company feeds you choose. Filter the rows and keep only job IDs that your monitor has not saved before.

Start with {"companies":["polestar"],"maxJobsPerCompany":10}. Open the Dataset for job rows and RUN_SUMMARY for empty, partial or failed results. Check the price shown before Start; no fixed selling rate is set by this guide.

Table of contents

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

What this Actor does

Add one to five public Teamtailor company slugs. A slug is the name before .teamtailor.com, such as polestar. The Actor reads each company's /jobs.json and /jobs.rss. It joins JSON items to RSS items by feed ID and saves one row for each matching posting ID.

Use a snapshot to get matching jobs in feed order. Use a monitor to get the newest matching IDs that have not been saved in that monitor. The first monitor run also returns current matches. Caps and filters leave unreturned jobs eligible for a later run.

This is an independent tool. It is not affiliated with, endorsed by or sponsored by Teamtailor. It reads public employer job feeds. It does not use a Teamtailor account key, read candidate profiles or submit applications. Teamtailor describes the public fields in its RSS feed guide.

What data you can get

Each job row has 20 keys. You can get the job ID, title, public description text, addresses, publication time and job link. Department and remote status come from RSS. Missing optional values stay null.

location is the first JSON address, with city, region and country when present. allLocations contains the unique addresses. RSS location names are used if JSON has no usable address. Location filters can match either source. country comes from the first JSON address; it stays null when that address has no two-letter country code.

The Dataset holds jobs. The default Key-value store holds INPUT, OUTPUT and RUN_SUMMARY. Monitoring adds a separate named store with saved job IDs. An empty Dataset is not enough to tell you why there are no rows: read the summary.

Quick start

1. Open this Actor's Input tab.
2. Switch to JSON input and paste the input below. No company is supplied by default.
3. Leave descriptions on. Set the output cap to 10 for your first check.
4. Click Start. Your Apify account must have access to the Actor.
5. Open the run's Dataset. Check company, jobId, title, location and url.
6. Open the run's Key-value store and check RUN_SUMMARY.outcome. Use companies[].errorCodes if the result is partial or empty.

{"companies":["polestar"],"includeDescription":true,"maxJobsPerCompany":10}

This input returned ten real rows in the saved example below. Live feeds can change. A cap of 10 limits saved matches, not the number of source items fetched.

To download results, open the Dataset and select Export, then JSON, CSV or Excel (XLSX) and Download. Leave the field selector empty to include all keys, or select your fields explicitly. Keep hidden metadata fields when you need _type or _scrapedAt. The default table view shows only nine columns; it is not the full field contract.

JSON keeps arrays and null values. CSV and XLSX expand arrays into columns such as allLocations/0; null cells can look blank. The saved ten-row Dataset was checked in all three formats. Use JSON if you need exact JSON types. See Dataset export options.

Input reference

All inputs are in one JSON object. companies is required. Omitted optional fields use the defaults below. A fresh Input form offers polestar, a two-job cap and snapshot mode (onlyNewSinceLastRun=false). You can run this sample without editing it; live job counts can change. API calls still require companies; there is no hidden fallback. Null is not an accepted input value. Unknown keys are rejected.

Key / form label	Type / required	Default / prefill	Allowed values	Effect and recommendation
companies / Teamtailor company slugs	string[] / yes	None / ["polestar"]	1–5 unique slugs; 1–63 letters, digits or hyphens; start with a letter or digit	Add known public boards, such as ["polestar"]. Values are trimmed and lowercased. Duplicate normalized slugs are rejected. URLs, custom domains and regional hosts are not accepted.
includeDescription / Include job descriptions	boolean / no	true / none	true, false	Use true for readable descriptions. False sets descriptionText to null; both feeds are still fetched.
maxJobsPerCompany / Job limit per company	integer / no	55 / 2	1–55	Limit matching output for each company. Start with 10. A snapshot takes feed order; a monitor takes newest unseen matches. A low cap does not make the feed request smaller.
titleKeywords / Words in job titles	string[] / no	[] / none	At most 20 values; each trimmed value 1–100 characters	Match any word or phrase as a case-insensitive title substring, such as ["engineer","intern"]. Empty means all titles.
locations / Locations to match	string[] / no	[] / none	At most 20 values; each trimmed value 1–100 characters	Match any value in JSON addresses or RSS names. Try ["Sweden"] or ["Göteborg"]. Empty means all locations.
remoteOnly / Only explicit remote jobs	boolean / no	false / none	true, false	True accepts only explicit RSS fully, remote, fully_remote, remote_first or remote_work. Hybrid, none and missing status do not match. Text mentioning remote work is not enough.
postedAfter / Earliest publication date	string / no	Omitted / none	YYYY-MM-DD, or ISO date-time with seconds and Z or an explicit UTC offset; optional 1–3 fractional digits	Inclusive publication-date filter. Example 2026-01-01 or 2026-01-01T00:00:00Z. A date-only value uses UTC. Omit it for all dates. Invalid dates and local times without a zone are rejected.
onlyNewSinceLastRun / Only IDs not saved before	boolean / no	false / false	true, false	False gives a snapshot. True emits and saves unseen matching IDs. First run emits current matches too. Use sequential runs, not overlapping runs.
monitorNamespace / Separate monitor name	string / no	"default" / none	1–64 letters, digits, underscores or hyphens; first character a letter or digit	Use a name such as my-sweden-watch. Keep it and the filters unchanged to continue a monitor. A new name starts a fresh baseline without deleting old state.

How settings work together

Settings	Effective behavior	What to do
Title values and location values	OR within each list; AND between the two lists	A job must match a title value and a location value when both lists are nonempty.
Filters and output cap	Filters run before the cap	Unmatched and unreturned IDs are not marked seen.
Monitor plus cap	Newest unseen matching jobs first	Repeat capped runs until the matching page is drained; later repeats can return zero.
Same monitor, new cap or description choice	The same saved-ID scope is used	Changing these does not reset the baseline.
New title/location/remote/date filters	A different state scope is used	Treat this as a separate baseline; filter order, case and repeated filter values are normalized for state identity.
Snapshot mode and namespace	No monitoring store is read or updated	Set onlyNewSinceLastRun true when you want persistent IDs.

Empty title/location lists disable those filters. Empty strings in those lists are rejected. The Actor does not search for every Teamtailor company.

Input examples

Minimal snapshot

{"companies":["polestar"]}

This uses the runtime cap of 55, descriptions on and no filters or monitoring. It returns available matching jobs from the first page. It is the smallest valid input, not a promise of 55 source jobs.

Small snapshot used for the output example

{"companies":["polestar"],"includeDescription":true,"maxJobsPerCompany":10}

This gives job rows in feed order. Check jobId, title, location, descriptionText and url. A source snapshot used with this input returned ten rows.

Separate Sweden monitor

{
  "companies": ["polestar"],
  "includeDescription": true,
  "maxJobsPerCompany": 2,
  "titleKeywords": [],
  "locations": ["sweden"],
  "remoteOnly": false,
  "postedAfter": "2026-01-01",
  "onlyNewSinceLastRun": true,
  "monitorNamespace": "my-sweden-watch"
}

This is a valid setup example. It starts a separate monitor for Sweden matches posted on or after that date. The first run emits up to two matches; later runs emit up to two unseen matches. It is not a claim that this exact namespace has already been run. Adding titleKeywords:["engineer"] narrows results further. Turning remoteOnly on excludes hybrid jobs and can make the Dataset empty.

Output example

This is one real row from the ten-row snapshot. All 20 keys are shown. Only descriptionText is shortened to a direct excerpt for this page; the actual Dataset contains the full text. The collection time is a sample time, not a freshness guarantee.

{
  "_type": "job",
  "_schemaVersion": "1.0",
  "_source": "teamtailor",
  "company": "polestar",
  "jobId": "3528756",
  "department": "Global Sales",
  "team": null,
  "title": "Internship | Retail Network",
  "location": "Cologne, Central Europe, DE",
  "allLocations": [
    "Cologne, Central Europe, DE"
  ],
  "workplaceType": null,
  "commitment": null,
  "country": "DE",
  "url": "https://polestar.teamtailor.com/jobs/3528756-internship-retail-network",
  "applyUrl": "https://polestar.teamtailor.com/jobs/3528756-internship-retail-network",
  "postedAt": "2024-01-17T14:42:18.000Z",
  "remoteStatus": "none",
  "updatedAt": null,
  "descriptionText": "The opportunity\nWe're currently looking for an Intern - ideally starting as soon as",
  "_scrapedAt": "2026-10-01T22:28:47.168Z"
}

The posting date can be much older than collection time. The Actor does not change publication dates to make jobs look new. A job is new to a monitor when its ID has not been emitted by that monitor, even if its posting date is old.

Output field reference

Job rows: all 20 Dataset fields

Field	Type / empty behavior	Meaning and source / sample
_type	string	Always job. This is a job row, not an error row.
_schemaVersion	string	Output contract version, currently 1.0.
_source	string	Always teamtailor.
company	string	Normalized input slug, such as polestar.
jobId	string	JobPosting identifier when supplied, then numeric ID in the validated job URL, then feed ID. Keep as text; sample 3528756.
title	string	Feed title with repeated whitespace collapsed; sample Internship | Retail Network. An item without a usable ID or title is skipped and reported.
department	string or null	Direct RSS department; sample Global Sales. Missing RSS value stays null.
team	null	No separate team value is extracted. It is not inferred from department.
location	string or null	First unique JSON address; RSS-name fallback only when JSON addresses are unavailable. Sample Cologne, Central Europe, DE.
allLocations	string[]	Unique addresses or fallback RSS names, in source order. No location gives [], not null.
workplaceType	null	No separate workplace-type classification is created.
commitment	null	No separate contract or employment-type value is extracted.
country	string or null	Two-letter code from the first JSON address, uppercased; sample DE. A missing first code stays null even if a later address has one.
url	string or null	Validated HTTPS job URL on the selected slug's .teamtailor.com host. Unsafe, redirected/custom-host or unsuitable values stay null.
applyUrl	string or null	Currently the same validated public job URL as url; not a separate direct application endpoint.
postedAt	string or null	Feed date_published, converted to UTC ISO time; sample 2024-01-17T14:42:18.000Z. Missing or invalid date stays null.
remoteStatus	string or null	Direct RSS label, such as none, hybrid or fully. It is not a guessed boolean.
updatedAt	null	No changed-job timestamp is extracted.
descriptionText	string or null	Readable text from content_text, then content_html, then summary. Markup is removed. Missing text or descriptions off gives null.
_scrapedAt	string	UTC collection time for this row; different from publication time.

Null means the output has no supported value, not that a job failed. Do not treat null as an empty string. A consumer that requires an always-null department needs to handle the RSS addition; this Actor does not promise that stricter contract.

Default Key-value store: OUTPUT and RUN_SUMMARY

INPUT holds the submitted settings with any defaults supplied by the platform. OUTPUT is a small result manifest. RUN_SUMMARY explains how the run ended. These records are separate from Dataset job rows.

OUTPUT field	Type	Meaning
schemaVersion	string	Manifest contract version 1.0.
status	string	succeeded, partial, not_found, failed or cancelled.
generatedAt	string	Summary finish time in UTC.
companyCount	number	Number of companies requested.
datasetItemCount	number	Number of rows acknowledged by Dataset writes.
companies	object[]	Per-company outcomes; all fields are listed below.
monitorState	object, optional	Present only in monitor mode; reference fields are below.

RUN_SUMMARY field	Type	Meaning
schemaVersion	string	Summary contract version 1.0.
outcome	string	Same semantic outcomes as OUTPUT status.
startedAt, finishedAt	string	UTC run-processing start and finish times.
durationMs	number	Elapsed product-processing milliseconds.
firstRowLatencyMs	number or null	Time until the first successful Dataset write; null when no row was written.
companiesRequested, companiesCompleted	number	Requested companies and companies that reached a non-cancelled result.
datasetRowsWritten	number	Acknowledged Dataset rows. Compare with Dataset count and OUTPUT count.
httpRequests, retries	number	Request attempts, including retries, and retry attempts. Two feed requests per healthy company are normal.
responseBytes	number	Bytes from successfully read feed bodies. Error bodies are excluded.
residentMemoryBytesAtFinish	number	Process memory observation at finish; not platform peak memory.
apifyComputeUnits, estimatedCostUsd	null	These are not calculated in the product summary. Check platform billing separately.
cancellation	string	none, requested, run_timeout, platform_abort or platform_migration.
companies	object[]	Per-company counters and errors.
fatalErrorCode	string, optional	Present on a fatal failure; may be invalid_input, actor_runtime_error, output_storage_error or summary_storage_error.
monitorState	object, optional	Present for monitoring; points to persistent state.

companies[] field	Type	Meaning
company	string	Company slug.
status	string	ok, partial, not_found, error or cancelled.
feedItemCount	number	Number of JSON items read, including rejected items.
validJobCount	number	Number of valid unique posting IDs.
duplicateJobCount	number	Valid duplicate posting IDs discarded.
emittedJobCount	number	Rows saved for this company after filters, seen-ID checks and cap.
rssMatchedCount	number	Unique valid jobs joined to RSS.
newJobBaselineCreated	boolean	True when this monitor/company had no previous state record.
errorCodes	string[]	Source, state or write errors; empty for a normal result.

Named monitoring state

OUTPUT and RUN_SUMMARY use the same monitorState reference:

Reference field	Type	Meaning
storeId	string	Apify ID of the named Key-value store.
storeName	string	teamtailor-monitor-v2- plus a scope hash.
namespace	string	Your monitorNamespace value.
schemaVersion	string	State format 2; separate from job contract 1.0.

Each company uses a SEEN-<UPPERCASE-SLUG> record, such as SEEN-POLESTAR.

State record field	Type	Meaning
schemaVersion	string	State format 2.
savedAt	string	UTC time of the last state write.
jobIds	string[]	IDs already saved to the Dataset by this monitor; up to 50,000 per company.

Only acknowledged rows enter the saved-ID set. State is isolated by Actor/account identity, namespace and title/location/remote/date filters; companies have separate record keys. Cap and description choice do not form part of that identity. Old v1 IDs are not imported into v2. No snapshot files, content diffs, changed/removed lists or history of every source job are produced.

Pricing and cost examples

Check the pricing model and charge unit shown on the Actor page before Start. For the pay-per-event configuration described here, a useful company feed check costs USD0.018 and each saved job row costs USD0.00045. A complete successful check with zero matches or no unseen IDs is still a feed check. This guide does not apply a pricing setting; the live Actor page governs your charges. Read the price and charge unit shown on the Actor page before Start. Use the run details for actual platform usage. Export and storage operations can also use platform resources.

Example	Scope / output	Cost and stop choices
Small snapshot	One slug, cap 10; the saved example has 10 rows	Check the current price. Set a timeout and the supported maximum charge limit before running.
Small new board	One slug, cap 3; a checked source gave 3 rows	Fewer saved rows do not guarantee a proportionally cheaper run: the feeds are still fetched.
Capped monitor	One slug, cap 2; checked repeats gave 2 then 2 distinct rows	Each repeat reads the feeds. Zero new rows can still use platform resources and incur the successful feed-check fee in pay-per-event mode.
Larger allowed input	Up to 5 slugs and cap 55 per slug	This is an input ceiling, not an observed price for 275 live rows. Do not estimate that bill from a ten-row example.

Descriptions off reduces output size, not feed fetching. Do not assume error or empty runs are free. See Apify pricing models and run charges.

Use cases

Buyer job	Input choice	Useful fields	Next step
Review an employer's current openings	Snapshot, one known slug	title, location, descriptionText, url	Export JSON or read the public job page.
Track new matching job IDs	Monitor, stable namespace and filters	company, jobId, postedAt, _scrapedAt	Add returned rows to your own database; deduplicate by company and ID.
Watch a location or role	locations and/or titleKeywords	allLocations, title, remoteStatus	Review the filtered rows and source links.
Check why a board is incomplete	Keep the run summary	Company status, errors, counts	Fix input or source access before relying on the result.

Scheduling and monitoring

1. Save a valid monitor input as an Apify Task if you want to reuse it.
2. Keep the same monitorNamespace and filters for that monitor.
3. Open Apify Schedules and choose that task or Actor with the saved input.
4. Choose the timezone and times you need. Allow enough time between runs to avoid overlap.
5. Check the first scheduled run's Dataset, OUTPUT and RUN_SUMMARY.

These are setup instructions; the Actor does not create schedules. See Actor and task schedules.

The first run emits current matching IDs and creates state. A checked capped monitor emitted two matches, then two different matches; its state grew from two to four IDs. Repeat until unseen matches are drained. Use a new namespace to reset logically; the old named store stays in your account. There is no reset switch or automatic deletion.

Keep monitoring runs sequential. Dataset writes and state writes are not one atomic transaction. If a state write fails, rows can reappear next time. An ID that changes its text or later disappears is not returned as a changed or removed job. Very old emitted IDs can fill the 50,000-ID state ceiling; choose a new namespace when you need a fresh baseline. An interrupted run is not automatically resumed as a content-history scan.

API, MCP and integrations

The API starts a run and returns a run ID and storage IDs. Poll that run until it reaches a terminal status, then read its Dataset. Use the build number your account can access; a tag is mutable. This example uses the exact private Actor ID. Set APIFY_BUILD to the immutable build number confirmed for your candidate; the saved output example comes from earlier build 0.3.1. Access must be granted to your account.

Set APIFY_TOKEN and the verified immutable APIFY_BUILD in your environment outside the script. Do not put a real token in this README, a URL or shared input.

curl --fail-with-body --request POST \
  --header "Authorization: Bearer ${APIFY_TOKEN}" \
  --header "Content-Type: application/json" \
  --data '{"companies":["polestar"],"includeDescription":true,"maxJobsPerCompany":10}' \
  "https://api.apify.com/v2/actors/wibBRjsLR3ahWeJNF/runs?build=${APIFY_BUILD}&memory=512&timeout=120&maxTotalChargeUsd=0.10&restartOnError=false"

Starting a run can cost money. This is a copyable example, not a run started by the guide. The maximum charge parameter is an Apify run option, not an Actor input field. Check the returned status and billing; do not retry an uncertain POST blindly. See the Run Actor API.

From the returned run, use defaultDatasetId with GET /v2/datasets/{datasetId}/items?format=json&clean=false. Use defaultKeyValueStoreId to read OUTPUT and RUN_SUMMARY. CSV and XLSX use the same items endpoint with format=csv or format=xlsx. Keep authorization in a header. Some Dataset access is possible by ID; do not share private run/storage IDs or exports with people who should not see them.

You can use the HTTP API from your own automation. Treat company + jobId as the job key, and check the summary before importing rows. This Actor does not provide its own MCP server or endpoint. No MCP service is installed by this guide. An agent must have a working, authorized API or platform connector before it can start runs; an MCP connection alone does not make this private Actor accessible.

Limits and partial results

• Use 1–5 unique public slugs on the global .teamtailor.com host. URLs, custom domains, .na.teamtailor.com and .au.teamtailor.com are outside the input contract.
• The output cap is 1–55 per company. Only the first JSON page is fetched. If it contains next_url, results remain usable but the company is partial with json_feed_paginated; the Actor does not follow that URL.
• Historical checks on build 0.3.1 emitted 10 snapshot rows and parsed a 55-job board for a three-row cap. Build 0.4.2 was checked with four small Cloud runs that emitted 2, 1, 1 and 1 rows. Ten-row comparisons and larger capacity evidence were reused from earlier checks. These are not proof of 275 live output rows or complete coverage of every employer.
• Each request has a 12-second timeout, at most one retry and a 5 MB response limit. Redirects are blocked. The Actor has a five-minute processing deadline; a shorter platform timeout or charge limit can stop it earlier.
• RSS may be absent, smaller than JSON or invalid. JSON jobs stay usable, but RSS fields can be null and the company is partial. An RSS-only location name or a remote filter can then exclude matches.
• The Actor does not infer salary, separate team, employment type, changed timestamps or remote status from prose. team, workplaceType, commitment and updatedAt remain null.

Situation	Dataset / summary	Meaning
Existing feed, no items or no matches	Empty Dataset; company ok, run succeeded	Valid empty result. Check filters or drained monitor state.
JSON feed returns 404	No rows for that company; not_found, feed_not_found	That public feed was not found. Not a timeout or valid empty board.
RSS unavailable or unmatched	Available JSON rows; partial, RSS error or rss_unmatched_items	Some enrichment is unavailable.
One company fails, another succeeds	Good rows retained; run can be partial	Read each company entry; do not discard valid rows from other boards.
Invalid input or technical failure	Failed summary when it can be saved; possibly no rows	Check the fatal error and platform logs. Do not label it a valid empty board.
Cancellation or storage failure	Previously written rows may remain	Inspect summary and platform status before using the run as complete.

Use public feeds with care. Public job data can change or disappear. Source availability and permissions are the employer's responsibility; you are responsible for your use and retention of the exports. No candidate data or private recruitment account access is part of this tool.

Troubleshooting

Symptom	Likely reason / error codes	Check and safe action
Input rejected	invalid_input; URL, duplicate slug, unsupported field, null or invalid date	Use the nine documented keys and a known bare slug. Check cap and date format.
No rows, normal outcome	Empty feed, filters or saved IDs	Check validJobCount, filters and emittedJobCount. For a fresh baseline use a new namespace.
Board not found	feed_not_found	Confirm the global public slug. A branded URL is not a substitute.
Feed cannot be read	timeout, network_error, rate_limited, http_5xx, http_4xx	Check the source later. Transient errors get at most one retry; do not create overlapping monitors.
Feed blocked or too large	redirect_blocked, response_too_large	Use a supported slug. The Actor will not follow an arbitrary host or raise the limit for this run.
Bad feed data	invalid_json_feed, invalid_json_feed_shape, invalid_job_items, rss_source_error	Read the source and summary; do not turn a parsing error into an empty-board claim.
Missing enrichment	RSS-prefixed request errors, rss_unmatched_items or rss_cancelled	JSON rows can still be useful. Check null values before filtering or importing.
Feed advertises more pages	json_feed_paginated	Treat the page as partial. Full pagination is not provided.
State cannot be read or saved	state_read_error, state_write_error	Check the referenced named store and access. Avoid overlap; state-write failure can cause repeats.
State is full	state_capacity_exceeded	The 50,000-ID ceiling stopped new emission. Choose a new namespace for a separate baseline.
Billing limit or problem	charge_limit_reached, billing_operation_error, billing_configuration_error, billing_duplicate_result	Read run billing and logs. The Actor checks the combined feed/row budget before writing. Unknown prices or synthetic extra fees stop work. It does not replay a nonempty run Dataset or retry an uncertain charge. Dataset delivery and event charging are not atomic; a billing failure can leave acknowledged rows.
Dataset/terminal records fail	dataset_write_error, output_storage_error, summary_storage_error, actor_runtime_error	Inspect platform logs. A missing record is not a successful empty result.

FAQ

Why can a monitor check cost money when no new jobs are returned?

A healthy check still reads the feeds and compares current IDs with saved IDs. In pay-per-event mode, the feed-check event covers this useful check. A duplicate or already-seen job is not charged as a new job row. Errors and partial checks with zero rows have no useful billable event; they can still use platform resources.

Do I need a Teamtailor API key?

No. The Actor reads public career feeds. You still need an Apify account with access and available usage for a Cloud run.

Does the first monitor run return zero jobs?

No. It returns current matching jobs up to the cap and saves only those IDs. Later runs return unseen matches. A low cap can give several nonempty repeats before zero.

Can I pass a custom career URL or find all Teamtailor employers?

No. Add known global Teamtailor slugs. The Actor does not discover all employers or resolve custom domains.

Why does an old job appear as new?

New means not previously emitted by that monitor. postedAt is the source publication date; _scrapedAt is collection time. Add postedAfter if you need a publication-date boundary.

Does hybrid count as remote?

No. remoteOnly requires an explicit accepted RSS value. Missing remote status and prose about flexible work are not enough.

Can I get changed, removed or every historical job?

No. State stores emitted IDs, not content snapshots or deletion events. Feed pagination, retention and employer changes can limit what you see.

Will a low cap or descriptions off cut the run cost by the same ratio?

No such cost reduction is promised. The feeds are still read. Check the shown price and the run's real usage; storage and exports can add resource use.

Related Actors

• Ashby Jobs Scraper — use the same owner's separate tool for public Ashby career boards.
• Lever Job Scraper — use the same owner's separate tool for public Lever boards.

These are separate products with their own inputs and prices. Their coverage does not expand this Actor's Teamtailor input contract.

Support

Open the Actor's Issues tab in Apify Console for a bug, question or feature request. Include the run ID, input mode, Dataset row count and relevant summary status/error code. Send only information you are allowed to share. Never include an API token, cookie, candidate record or private billing payload.

The examples and artwork explain the public job-feed workflow. They do not imply Teamtailor endorsement, a full job-history service or a source availability guarantee.