How duplicate jobs are detected

We flag jobs as duplicates to keep search results clean and useful for candidates. This page explains what we look at and what you can do about it.

How the check works

When a job is created or edited, we generate a vector embedding from its description and compare it against your other recent jobs and the jobs of your organisation. If the similarity score is above the configured threshold, the newer job is marked as a duplicate and hidden from search results.

What is taken into account

The full text of the job description (HTML tags are stripped before comparison).
Whether jobs belong to the same user or the same organisation — same user matches use a stricter threshold.
Only jobs with at least 20 characters of description are checked.
Premium and inactive jobs are excluded from the duplicate index.

Title, city, salary and tags are not used directly for the duplicate score — they are stored alongside the embedding for display only.

How to avoid duplicates

Move shared boilerplate (about-us, benefits, hiring process) out of the description.
Add company-wide content once on your profile page so it does not need to be repeated on every job.
Make each description focus on what is unique to the role — responsibilities, requirements, the team.

Common parts that apply to every job (about the company, perks, benefits, recruitment policy) should live on your profile, in the section linked below. They will then appear automatically next to every job and won't trigger duplicate detection.

Open profile → Terms and conditions

What to do if a job was flagged

Open the duplicate badge in the job list. You can either promote the job (it becomes visible again) or edit it — five minutes after posting, the Edit option becomes available. Saving an edit clears the duplicate flag and runs a fresh check.