What Legora Costs a UK Law Firm, and What You Get for It

Search for Legora's pricing and the pages I found were all written by rival vendors, most of them in the United States. That's not a criticism of them; it's a gap. I haven't found it written up for a UK firm from the buying side, so this is that piece. Every number below is attributed, because none of them come from Legora, and none of them should be read as Legora's own price.

What the public numbers say

I couldn't find a public Legora price list; legora.com has no pricing page. What exists is two rival vendors' estimates, and both say so. Spellbook's comparison, updated at the end of August 2026, estimates Legora at about $3,000 per user a year with a 10-seat minimum, and labels both figures as estimates. Vaquill's pricing round-up, updated in July, reports a wider range of $300 to $800 per user per month depending on scale and a minimum annual contract of around $30,000, and says the Agent Pro tier moved to consumption-based pricing in June 2026; I haven't had that confirmed by Legora. Vaquill also reports buyers securing 40 to 60 per cent off the first quote. Treat all of that as third-party estimates rather than Legora's own pricing, but the recurring claims are consistent: a roughly ten-seat minimum, an annual minimum in the low thirties of thousands of dollars, and possible usage charges on top for heavy agent work.

For scale, Vaquill reports Harvey at $1,200 to $2,000 or more per seat per month with a 25-seat minimum, which it estimates as a $360,000 annual floor before anyone has logged in. On those figures Harvey isn't a like-for-like option for most UK firms; it's a different league of buyer. On the same estimates, Legora is the one a mid-sized UK firm can realistically have a conversation about.

What you get for it

The product is a layer over other companies' models rather than a model of its own. Anthropic's own customer page says Legora uses Claude across its assistant tools, document review and workflows, and earlier reporting had it running OpenAI and Meta models through Microsoft Azure, which is the substance of the wrapper question I wrote about in July. After OpenAI released GPT-6 Astra in early September, Legora described its own test of the model in Artificial Lawyer on 7 September: it said its agent completed a financial-statement tie-out across 41 documents in a single run and found all four errors the team had planted, including a £500,000 gap. That's Legora marking its own homework, but it tells you what the product is now built to do, which is finish multi-step jobs rather than answer prompts.

Around that sit the things a firm notices day to day: Tabular Review, which turns a folder of documents into a grid with a row per document and a column per question, a Word add-in for drafting, and the integrations, with iManage established and NetDocuments, Box, Intapp and Avvoka announced in a rush through August. If your document management and time recording are on any of those, the integration is something to test in the demo rather than take on trust.

The newest piece is research. After buying Qura in April and Wexler in July, Legora announced on 14 September a full ontology of law and an AI-native citator; Artificial Lawyer reports the ontology is in limited beta with general availability due in the fourth quarter. Arvid Winterfeldt, Qura's co-founder and chief executive, who now leads legal research at Legora, put it this way: “We have catalogued more than 50 distinct ways AI fails at legal research.” The August data announcement covered US primary law. For a UK firm the question to ask is how much of England and Wales the citator will cover at launch, and whether it's inside the seat price or on top.

Who in the UK has bought it

Reports of the $550 million Series D in March, which valued the company at $5.55 billion, named Linklaters and Bird & Bird among its customers. The more useful signal for most readers came this month, when Edwin Coe, a Lincoln's Inn firm with £44 million of revenue in 2025/26 and more than 250 employees, announced a firm-wide rollout across all 14 of its practice groups after a five-week pilot, live since July. A magic circle customer name tells you the product survived procurement. A mid-size London firm putting it in front of every practice group after a five-week pilot is stronger evidence of adoption than a logo.

What the headline price hides

Vaquill's estimates for the category, not Legora specifically, are worth having on the table when the quote arrives: implementation and onboarding anywhere from $5,000 to $100,000, training at $500 to $2,000 a user, premium support at 15 to 20 per cent of the licence, and a year-one total 30 to 50 per cent above the headline. It also reports renewal uplifts of 10 to 25 per cent a year, often uncapped, on Harvey contracts. Whether Legora's terms carry the same, I don't know, and neither will you until you ask for the escalator in writing.

The reported consumption pricing on the agent tier deserves its own line. A tie-out across 41 documents is the kind of workflow where you need to know how usage is measured, because if your associates come to trust it they'll run it every week. Ask what a credit is, what a typical review costs in credits, and whether you can cap monthly spend per team.

The questions to put in writing

• Which model handles each workflow, and where is the data processed? Legora is a Stockholm company; confirm the contracting entity, the processing region and the retention terms for your client documents.
• What counts as a seat? Fee-earners only, or everyone who opens a document?
• What does the citator cover for England and Wales at general availability, and is it in the price?
• What's the unit of consumption on the agent tier, and can you cap it?
• What's the renewal escalator, and is it capped?
• What leaves with you at exit: the workflows, the review grids, the prompts your team built?

What do they build that you couldn't?

Put that question to the vendor, because the honest answer is shorter than the demo. The models are rented; you can rent the same ones through Amazon Bedrock or Azure. The security layer, the document handling, the Word plug-in and the integrations are real engineering, but a firm with a capable IT team can now buy most of it in pieces. What you can't easily assemble yourself is the research layer: licensed primary law, structured and maintained, with a citator on top. That is the scarce part of what Legora sells, and it's the part to weight when you compare the seat price with building around the raw model. If the answer to 'what can't we build' turns out to be 'the citator', then you're paying for the citator and everything else is convenience.

If AI stopped tomorrow

One more question, and this one is for the partners rather than the vendor. If the tool went dark tomorrow, through an outage, a renewal you can't stomach, a model provider changing its terms or a regulator stepping in, could the firm still run? Could your associates still do a tie-out, and would anyone below partner still know how? The consumption tier, the multi-year terms and the escalators all assume you'll be more dependent in year three than in year one, and the pricing is built on that assumption.

Ethical AI in a law firm covers bias and confidentiality, and it also covers something plainer: keeping the firm's judgement inside the firm. Supervision that's real rather than nominal, clients told what the tool did on their matter, and a business that would survive the tool being switched off. Buy it if the numbers work. Don't let it become the only place the knowledge lives.

The part the price doesn't settle

Whatever Legora costs, every firm that licenses it gets the same core product. A shared tool can't be an edge; at best it's the cost of staying level, which I argued in AI won't differentiate your law firm. The money question is the easy one, and the vendor will answer it in a demo. The harder one is what you'll build around the tool that the firm across the road won't, and that's the one worth a conversation before you sign anything. If you're working through an evaluation now, the practical guide to assessing legal AI for UK firms is the place to start.