This week Archivers.ai launches commercially — and with it, we are doing something the sector’s software market has never done. The archives sector makes its case for openness every day: open records, open access, accountability through documentation. Launch week seemed like the right moment to apply the same standard to ourselves.
As of this week, every price Archivers.ai charges is published on our pricing page — every plan from free to Institution, every credit pack, every per-unit cost, every footnote about VAT and credit expiry. No “contact us for a quote”. No pricing that depends on who is asking. As far as we can tell, these are the first published AI cataloguing prices anywhere in the sector.
This post is about why we did it, why almost nobody else does, and what it changes for the people who actually have to get cataloguing projects funded and approved.
The quote-only convention
If you have ever tried to budget for collections software, you know the routine. The website describes the product warmly and at length, and then the pricing page — if there is one — says contact us. So you book a call. You describe your service, your collection size, your user numbers. A proposal arrives some weeks later, marked commercial-in-confidence, and you have no way of knowing whether the figure in it is what the next archive was quoted, or double.
To be clear: this is the convention, not a conspiracy. Quote-only pricing is the norm across enterprise software generally, and there are defensible reasons for it — implementations vary, support needs vary, large deployments genuinely are bespoke. We are not suggesting other vendors are doing anything improper. But a convention designed for corporate IT procurement lands very differently on a sector made up of small, publicly accountable, grant-dependent organisations. What works for a multinational negotiating an ERP contract actively hurts a two-person record office trying to cost a cataloguing project.
What hidden pricing costs an archive
The costs of opacity are specific and practical.
It makes grant budgets guesswork. NLHF and Archives Revealed applications need costed budgets, often a year or more before any money is spent. If the software line in your bid depends on a confidential quote, you either start a sales process just to write an application that may not succeed, or you put in an estimate and hope. A published unit price removes the problem: the figure goes straight into the budget table, with a public URL as its source. An assessor can check it. So can your committee. (Our heritage funding guide covers how to structure that budget line.)
It wastes the scarcest resource: staff time. Discovery calls, demos scheduled three weeks out, proposals to chase — all to learn a number. Archivists tell us the evaluation process for collections software is measured in months. Most of that time is spent extracting information that could simply have been on a webpage.
It tilts the table. When only the vendor knows what everyone pays, every negotiation starts unequal. Small services with no procurement department are least equipped to push back — and they are most of the sector.
It quietly excludes the smallest archives. A volunteer-run community archive will often not book a sales call at all, on the reasonable assumption that “contact us” means “more than we have”. Quote-only pricing filters out exactly the organisations with the largest uncatalogued backlogs and the least capacity to navigate procurement.
Transparency is a trust position, not a pricing tactic
We are asking archives to do something genuinely serious: to let AI assist with catalogue records — the evidence layer of their institutions. Our answer to that seriousness is a trust architecture we have written about at length: nothing enters the catalogue without human review, every AI-suggested field cites its evidence, and provenance for AI-assisted fields — model, timestamp, review state — travels with the record. The full picture is on our trust page.
Published pricing belongs to the same architecture. A platform whose entire pitch is AI that shows its working cannot then hide its own commercial working behind a sales process. If we expect archivists to trust what the system says about a photograph, the least we can do is be equally inspectable about what the system costs. Trust is not divisible — you either operate in the open or you do not.
There is also a plainer point. Our pricing can survive being public. Most customers pay 7–10p per catalogued item or page, against the £3–5 per item that manual cataloguing costs in grant-funded projects — we have set out the full cost arithmetic, sector backlog figures and a worked 7,000-item example separately. When the comparison is that stark, publication is not brave; it is just the obvious move. The interesting question is what it tells you when a price cannot survive publication.
What published prices make possible
The practical payoff shows up everywhere an archive has to justify spending.
Straight into the grant budget. A published price list with public unit costs — 10p a credit on a 10,000 pack, 8p on the 50,000 pack we sized for an Archives Revealed cataloguing grant — can be cited directly in an NLHF or Archives Revealed budget, with no commercial-in-confidence caveats and nothing for an assessor to query. Customers fund Archivers.ai through project budgets today, and the published page is what makes that clean.
Straight into the committee paper. Budget-holders can put our prices in front of a finance committee, a board, or a section 151 officer without redactions or “indicative figures subject to final quotation”. The paper says £320 a month, annual, and anyone who doubts it can look it up.
Simpler procurement. Most of our plans sit well under standard purchasing thresholds, which means a card or a purchase order rather than a tender. Where rules require three quotes for larger purchases, a published price gives your procurement team a fixed, verifiable benchmark from the start. We take invoices and purchase orders for Institution plans and packs of £1,000 and over.
Honest self-qualification. Perhaps most importantly: you can rule us in or out in ninety seconds, on your own, without talking to anyone. A community archive can see that the free plan costs nothing and the Starter plan is £40 a month. A county service can see the Institution tier at £400 a month and know roughly what a backlog project would add in credit packs. Nobody’s time is wasted on a conversation the website could have settled.
The commitments behind the page
Publishing prices once is easy; the commitment is to how the page behaves over time. So, plainly:
- What you see is what everyone sees. The published price is the price. We do not run a different rate card for organisations that negotiate harder.
- No surprise arithmetic. Credits are defined publicly — one credit catalogues one photograph or object, one document page, one minute of video, or five minutes of audio — and the app shows the credit cost of any job before you commit. Monthly credits reset; top-up packs last 12 months; prices exclude VAT. All of it is written down.
- Custom means custom scope, not custom secrecy. Enterprise and consortium arrangements are genuinely bespoke — pooled credits, multiple branches, custom storage — but they are built openly from the same published unit economics you can already see.
We will not pretend any of this is costless to us. Published prices can be undercut, quoted against us, and held against us if we ever change them. We think that discipline is the point. A sector whose entire professional ethic is documentation and accountability deserves suppliers who can live by the same standard.
An experiment in openness
Archives exist because institutions are healthier when their workings are inspectable. We have simply applied the principle to ourselves. Whether the rest of the sector’s software market follows is not ours to decide — but every archive that walks into a procurement conversation holding a public benchmark makes the old routine a little harder to sustain.
The prices are on the pricing page. The cost arithmetic — manual £3–5 per item versus published per-credit pricing, with the full worked example — is in our companion post. And the fastest way to evaluate us costs exactly what the page says it does: start free, no card, 30 credits a month, and see what your own backlog looks like with a price on it.