Spellbook

What it actually does

Spellbook is a Microsoft Word add-in for contract work. It reviews a draft, flags risk, suggests alternative language, produces redlines with real-time suggestions, and benchmarks terms against market standards. Firms can configure playbooks so the review reflects their own positions rather than a generic standard.

Running inside Word is the right design decision and the main reason to consider it. Transactional lawyers already live in Word, and a tool that works in the document beats one that makes you upload a contract to a separate platform and bring the output back. It is not a legal research product, so it does not compete with the broader legal AI suites.

What it costs

We cannot tell you, and neither can Spellbook's website. There is no published price list anywhere on the site, so learning the cost requires booking a sales call. For a product sold to lawyers, who negotiate for a living, hiding the number is a strange opening move.

A 7-day free trial with full access is advertised. Reviewers say they could not actually get one without going through sales first, which makes the trial a lead magnet rather than a trial. One Trustpilot reviewer states plainly that the company claims to give a free trial but never does.

The brand also splits across two domains, spellbook.legal and spellbook.com, which does not help a buyer work out what they are looking at.

Where it falls down

The public record is small and it is not good. Spellbook holds 2.7 out of 5 on Trustpilot from 10 reviews, with 30% at one star and 40% at five, and only two of those reviews landed in the last twelve months. Trustpilot notes the company has not replied to its negative reviews. Ten reviews is a thin base and we are not going to pretend it is statistically meaningful, but it is the only independent record that exists, and a vendor that ignores every negative review on a ten-review profile has made a choice about how it handles criticism.

The homepage advertises 4.7 on G2. We could not verify that on G2 itself, so as far as this page is concerned it is a marketing claim printed by the vendor, not an independent rating. Treat it that way until someone can open the underlying profile.

The conduct complaints are specific. One reviewer describes an unrequested cold call and being hung up on. On output, lawyers in the review set dispute the quality: one says the drafting is bad enough to waste billable hours, and another says general-purpose Claude does the job better. A separate reviewer offers the opinion that the legal AI positioning could expose the vendor to misrepresentation claims. That is an opinion, not a filed action, and we found no lawsuit or regulatory action against the company.

Set against that, the positive reviewers are consistent about what they like: the native Word workflow and, from earlier adopters, transparency on pricing. Legal tech commenters on Reddit reportedly call it the best drafting and redline tool they have used, though Reddit blocked extraction so we could not verify those posts directly.

Who should buy it

A firm or in-house team whose work is overwhelmingly contract drafting and review, that already runs everything through Word, and that has the buying discipline to handle a sales-gated process. If that is you, take the call, insist on a working trial on your own documents before any commitment, and get the renewal terms in writing.

Do not buy on the demo. Two lawyers on the public record say the output cost them time, so the only evidence that matters is your own contracts running through it, and you should not sign anything until you have that.

Do not buy it for legal research. It does not do that.

And if you are a solo practitioner or a small team, consider whether a general assistant does enough of this for a published monthly price. One reviewer reached that conclusion already. Two cyders reflects a product that may well work, sold in a way that gives a buyer almost nothing to verify before paying.