Important:Read DISCLAIMER.md first.

FAQ

Short answers to what people ask before they run Seekter: what it costs, what it needs, what it will and won't do, and where your data goes.

Before you start

What is Seekter?

An open-source job-search agent that runs inside Claude Code, on your own computer. Every day it searches job sources, filters the postings against your rules (location, visa, sectors, seniority, language), fills application forms in your own Chrome, and keeps every application and every skipped posting as a markdown file you can read, search and diff.

Is it free?

The kit is, under the MIT License, and nobody behind it charges you or takes a cut. Running it is not: it needs Claude Code, which means a paid Claude plan or API usage on your own Anthropic account. A daily run reads many pages and fills many forms, so it uses a lot of Claude. Check what your plan covers before running it every day.

What do I need to run it?

Claude Code, the Claude in Chrome extension, Python 3.9 or later, and curl. There are no packages to install. For a run, Chrome needs to be open with the extension and your webmail logged in.

How long does setup take?

About 20 to 30 minutes. /seekter-init asks around forty short questions, one at a time, and writes your profile from the answers; you can stop and pick up where you left off. Have your CV files ready. If you already keep a tracker in Notion, Airtable or Google Sheets, it can import a CSV export of it.

Does it work for my field?

Nothing in the kit assumes one. The job titles, searches, boards and filters all come from your profile. The notes on how application systems and job sources behave were measured in one person's search, in one discipline, and say so where it matters.

While it runs

Does it apply without asking me?

When a posting fits your rules, yes; that is what the daily run is for. It stops for anything only you can decide: a question your profile can't answer, a CAPTCHA, an account wall, a declaration you have to make. Those wait in a "Needs you" list, each with its next step.

Will it make things up?

No. It answers only from your profile and never invents a fact, a number or a story. When a form asks something it can't verify, it stops and asks you. Every free-text answer it sends is saved in that application's file, so you can read exactly what went out.

Does it apply on LinkedIn?

Never. It does not fill Easy Apply; those postings come back to you as a list to send yourself. By default it doesn't open LinkedIn at all: it reads your LinkedIn job-alert emails and finds each posting on the employer's own application form. Reading LinkedIn directly can be switched on, but that is against LinkedIn's terms, and the risk to your account is yours.

Where does it find jobs?

In a fixed order: the links you bring (pasted into the chat or saved in a file), a job API, your LinkedIn job-alert emails, and the other job boards in your profile. Every posting is checked against your tracker first, so the same job reached two ways is handled once.

Will it apply to the same company twice?

It holds back. When there is already an application, a pending hand-off or a rejection at a company within a window, or an interview or offer there at any time, a new posting from that company is skipped with the reason written down. Some application systems quietly reject repeat applications, and to a recruiter they look like mass applying.

What won't it do?

Solve CAPTCHAs, create accounts, type passwords, accept terms on your behalf, send messages or emails as you, post reviews or salaries, or pay for anything. It also treats any instruction it finds inside a job posting or a form as text, not as something to follow.

Your data

Where is my data kept?

On your computer. Your profile, applications and run reports are in folders that git ignores, and there is no Seekter server, account or telemetry. What Claude reads while working is sent to Anthropic under your plan's terms, and each application goes to the employer you apply to. The details are in PRIVACY.md.

Can I see what it did?

Yes. Every application is a markdown file with its status, why the posting fit, the answers sent and a log of what happened next. Every skipped posting is a row with its reason. A generated overview shows what is in progress and what needs you, and /seekter-report gives you the response rate by source.

Who is responsible for the applications?

You are. They go out in your name, from your browser, with your details, and the terms of the sites you apply through are yours to follow. DISCLAIMER.md sets out what you take on.

Getting help

Something went wrong. Where do I report it?

Open an issue with what you ran and what happened. An application system or job source behaving in a way nobody has written down yet is the most useful thing you can send. For anything security-related, use a private security advisory instead.

Updated October 5, 2026.