Everything people ask us.

The slot is released immediately and you can publish the next one in its place. Slots are concurrent, not cumulative: five roles on Team means five open at once, with no cap on how many pass through in a year.

Yes — $240 a month per extra concurrent role, on any plan, as many as you want, added or removed whenever you need them.

Larger plans cost less per role, so past a certain point upgrading is simply cheaper. We will tell you when you get there rather than let you overpay.

Not from our pool. No count, no threshold, no overage, no credits to top up. A busy month and a quiet month cost the same.

Applicants you bring us yourself are different — you control that volume, so those are priced per CV on import. It is the principle the whole price list rests on: we charge for what you control, and never for what you don’t.

Your first role, free, if you claim it before 31 December 2026. The date is the deadline to claim, not an expiry on the role: once it is yours, it stays free until you fill it or close it.

We are still building the candidate pool, so early customers do not pay until it is there. One free role per company, and a twelve-month ceiling.

$750 covers the first 3,000 CVs, then $0.18 each above that. It is charged once, on import — a one-time job, not a subscription, and nothing recurring is added to your plan.

Setting the roles up is part of onboarding, so there is nothing to pay for that.

More than 10 roles, or more than 10,000 CVs, and we scope it with you before anything moves rather than quoting you a number that turns out to be wrong.

No. Re-runs are unlimited on every plan, including the free one, and re-scoring an imported candidate is never charged twice.

That is deliberate rather than generous: every time you adjust the weights and tell us the list got better or worse, our scoring gets more accurate. We would rather you iterate.

No. There is a fair-use guardrail — twice your slot count per month, minimum three — and it exists only to stop someone cycling roles to sit in a lower band.

It cannot fire on normal hiring. If you ever came close to it we would talk to you, not bill you.

No — a subscription covers hiring by the contracting company. Agencies, search firms and multi-entity groups need a separate agreement. Talk to us and we will sort out something that fits.

You talk to us. Every account is opened by hand at the moment: we set the role up with you, agree the weights, and stay on it. It is slower than a signup form, and it is the reason early customers get a shortlist worth reading.

Monthly or annually, invoiced, in USD. Annual billing starts at Team. Nothing is charged before we have set you up.

You build it as structured requirements — skills and the level you need them at, years of experience, education, and the weight you put on each — through a guided form. The written job description is generated from that structure.

We deliberately do not work the other way round. Pasting prose and having a model guess your requirements reintroduces exactly the ambiguity that makes CV screening unreliable in the first place.

You do. Always. Weights are never inferred by a model — if skills matter three times as much as education for this role, you say so, and the score reflects exactly that.

This is the part most tools get backwards. A system that decides for itself what matters cannot be argued with, and cannot be audited.

The score a candidate has to reach before they appear on your list for that role. You set it per role — the default is 60 — and you can move it whenever you like.

It is the one dial that controls how much you see. Raise it for a flooded role, lower it for a scarce one.

Because names, schools and previous employers are the things that produce biased first screens, and hiding them is the whole point. Until you shortlist someone you see the profile and the score breakdown, not the person.

You are choosing on evidence you can defend. That is the argument, and it only works if we hold the line on it.

That is the moment of introduction. Their identity unlocks for you, and on their side the full job description and your company name unlock at the same time.

Until then the candidate sees only a title and a score — they do not know it is you, and you do not know it is them. Shortlisting is both of you agreeing to find out.

Yes, and it stays published — there is no path back to draft.

Editing does not silently re-do anything. You get two separate buttons: regenerate the description, and re-run matching. They are independent because you often want one without the other, and because a system that re-ranks your list without being asked is a system you stop trusting.

No. Closed is final — to revive a role you create a new one.

It sounds strict, and it prevents a specific unfairness: a candidate who was told a role was closed should not silently find themselves back in a live pipeline months later without knowing.

Never. Rigel does not post anywhere on your behalf, and it is not on the roadmap.

What you get is an apply link. Paste it wherever you are already advertising, and everyone who follows it gets scored and gets an answer — which is more than most application forms manage.

The title and your score. Not the company, not the full description.

That is the same wall the recruiter is behind — they cannot see your name either. When they shortlist you, both sides open at once: you get the company and the full description, they get your identity.

Recruiters whose role you scored well enough for, and only in anonymised form until one of them shortlists you. Nobody browses candidates.

The detail on how your data is stored and handled is in our privacy policy, and we would rather you read it than take a reassuring sentence from an FAQ.

Adding skills you actually have will. Adding skills you don’t have will raise a number and cost you the interview — the recruiter sees the breakdown behind your score, so an inflated profile fails in the room instead of on the page.

Rigel is built to reflect your current profile, not to be a target you optimise against. A platform that rewarded padding would be the thing we set out to replace.

Yes — one toggle in your settings turns matching off. You stop being scored against new roles immediately, and you stay off until you turn it back on.

Useful when you are not looking, or when you have accepted something and want to go quiet without deleting anything.

Yes. Any match you have been shortlisted for has a decline option, and declining removes you from that role’s pipeline. No explanation needed.

If someone sends you a Rigel apply link you can use it directly, without an account: upload your CV, check what we read from it, and get your score for that role.

If you make an account later with the same email, that profile carries over and starts being matched everywhere else. You do not start again.

Yes, whenever you like. Every confirmed version is kept, and re-confirming re-scores you against every open role — so a new job, a new certification or a skill you forgot takes effect straight away.

The reading is. A model extracts structure from a CV — roles, dates, skills, education — because CVs are prose and prose needs parsing. It also drafts the written job description from the recruiter’s structured requirements.

The score is arithmetic. Once the CV is structured, matching is the weighted sum above. No model decides how well anyone matches.

Yes, all of it. Every score opens into its sections, every section into the requirements it was measured against, and every run stores the weights and the scoring version used at the time.

That last part matters more than it sounds: it means a score from three months ago can be reproduced exactly, which is what “auditable” has to mean to be worth anything.

No. We do not train, tune or validate models on customer data or on applicants a customer brings us.

If we ever wanted to, we would ask you separately and specifically, and you would be free to say no. It would not arrive inside a broad clause in a contract you already signed.

Not yet, and we would rather say so plainly than imply otherwise.

What we do have is the thing an audit needs: scores that decompose, weights set by your recruiter rather than inferred by a model, a first screen with no name or school in it, and every run stored with the weights and version it used. An independent auditor can get real numbers out of Rigel, which is more than can be said for a black box.

If you want to run one during a pilot, we will support it.

No. Rigel produces a ranking and the reasoning behind it. Every decision to advance, shortlist or reject is made by a person, and the product is built so that person has something real to base it on.

Recruitment tools are high-risk under Annex III of the EU AI Act, and those obligations apply from 2 December 2027 after the Digital Omnibus deferral. We are building to be conformant ahead of that date.

What applies today is GDPR — including Article 22, which says a decision with significant effects cannot be based solely on automated processing. Rigel is designed for that rather than around it: a human decides, the weights are the employer’s, and every score can be explained to the person it is about.

EU customers are hosted in the EU. For anything more specific, ask us — we would rather answer properly than in a paragraph.

Candidates own their profile, can turn matching off, and can update or withdraw it.

Applicants a company imports are handled under that company’s agreement.

The specifics — what is stored, for how long, who processes it — are in the privacy policy and, for customers, in the data-processing agreement. Those are the documents that bind us, so those are the ones worth reading.