The Westlaw alternative.
Westlaw is the best research product ever built for a specific kind of practice. If you are not that practice, you pay for a citator you touch twice a month.
This page names which practice Westlaw is right for, prices the alternatives with dated sources, and says where Marcella does not compete.
Prices verified 2026-08-24. Every figure carries its source and read date.
Where Westlaw is still right
Starting here, because the rest of the page is worthless if this part is dishonest.
You KeyCite daily
If citation verification is a repeated daily motion across a litigation team, the KeyCite editorial layer and West headnotes are the product. Four decades of human classification is not reproduced in a release cycle.
You live in secondary sources
Practical Law and the West treatise library are deep and exclusive. If your research starts in a treatise rather than a case, no alternative on this page replaces that.
Opposing counsel expects it
In some practice areas the West citation format and the analytics are the shared language. That is a real switching cost and it belongs in the decision.
The case against switching is the strongest part of this page. If any of the three describe your practice, nothing below changes your renewal. Read the pricing breakdown and negotiate the term instead.
The published numbers
Thomson Reuters changed posture in 2026 and now publishes self-serve prices for new-customer firms with up to ten attorneys. Eleven or more still receive a quote. That change is what makes an honest comparison table possible for the first time.
| Product | Single circuit or state | All states plus federal | Notes |
|---|---|---|---|
| Westlaw Edge | $107.25 / mo | Not published | Three year term. AI citator plus litigation analytics.Thomson Reuters pricing pages via Spellbook, 2026 |
| Westlaw Edge with AI-Assisted Research | $155.35 / mo | $266.50 / mo | Three year term.Thomson Reuters pricing pages, June 2026 |
| Westlaw Advantage | $256.75 / mo | $399.75 / mo | Current flagship, launched August 2025. Folds in what was sold as Westlaw Precision and CoCounsel branding.Thomson Reuters, June 2026 |
| Westlaw Classic | $133 / mo | $266.40 / mo | One year commit. All states without federal reported at $200 per month.Lawyerist, December 2024 |
| LexisNexis entry | $114 / mo | $200 to $300 / mo | Solo plan entry rate. Lexis+ AI at enterprise reported near $17,500 per year.thelawgpt.com survey, June 2026 |
| Marcella Firm | $99 / seat / mo | $99 / seat / mo | 1 to 25 seats, annual billing, Technology Outlaws hosted. No jurisdiction tiering. Launch pricing. |
| Marcella BOYA | $278 / mo flat | $278 / mo flat | 3 seats. Deploys in the client's own Azure tenant. Launch pricing. |
Read the term, not the rate. The published Westlaw AI rates are three year commitments. A two year contract carries a reported twelve percent discount and three years carries eighteen percent, which means the headline number already assumes you signed for three years. Price the one year equivalent before comparing anything.
What the sticker price leaves out
Out of plan charges
Content outside your selected plan is billed per search and per document. The Franklin County Law Library published rates of roughly ninety nine dollars per Westlaw search and roughly sixty dollars per Lexis search for out of plan use, updated June 16, 2026. An associate pulling twenty documents in one session accumulates these quietly.
Retraining on every rebrand
Westlaw Edge, then Westlaw Precision, then Westlaw Advantage with CoCounsel folded in, inside roughly three years. Each transition consumes lawyer hours that are not billable and do not appear on the renewal.
The silo tax
Each platform searches its own content. Westlaw searches Westlaw. Lexis searches Lexis. Clio acquired vLex for one billion dollars in November 2025 and Clio users search vLex, without Prosser, Wright and Miller, or Collier unless they subscribe separately. Firms that need coverage across publishers pay more than once. Full breakdown.
The reliability question
A Stanford RegLab and Yale study of leading AI legal research tools, published in 2024, tested the major paid products for fabricated or unsupported output. The reported finding was that the Westlaw AI-assisted research product produced inaccurate output in roughly one of three responses, the weakest result among the paid tools tested. Lexis performed better and still produced hallucinated content at a measurable rate.
That study is why retrieval architecture matters more than model quality in this category. A generative system asked for a citation can produce a plausible one whether or not the case exists.
Marcella's constraint sits below the model. Citations are retrieved from an indexed corpus of more than 10 million opinions. Not in the index, not in the output. A fabricated citation is not a probability that has been reduced. It is a result the retrieval path cannot produce. Six-item Daubert attestation on every citation.
The running record of what happens in court when that constraint is absent is maintained at heppner-problem.com, an independent reference site. As of mid 2026 the underlying public database tracked more than 1,500 court decisions worldwide involving AI generated material, and penalties have moved from four figure fines to license suspensions.
Feature by feature
| Capability | Westlaw | Marcella |
|---|---|---|
| Editorial citator with treatment classification | KeyCite. Four decades of editorial depth. The category standard. | Citation treatment engine mapping the public record across publishers, with linked citing passages. |
| Secondary sources and treatises | Practical Law and the West treatise library. Not replicated elsewhere. | Not a treatise publisher. Treatment coverage spans 26 canonical treatise families, which is a different thing than licensing the text. |
| Fabricated citations | Generative layer. Measured failure rate in the Stanford and Yale study. | Retrieval only. A citation absent from the index cannot be returned. |
| Where client data sits | Thomson Reuters infrastructure under the subscription terms. | The firm's own Microsoft 365 tenant. On BOYA tiers, the firm's own Azure tenant. No external AI vendor in the data path. |
| Where you use it | Browser and dedicated application. | Inside the Microsoft 365 tenant the firm already runs. No new vendor approval. |
| Pricing transparency | Published up to ten attorneys. Quote above that. | Published at every tier through BOYA 20. |
| Cross-publisher coverage | Westlaw content. Other publishers require other subscriptions. | Citation relationships mapped from the public record, which crosses every publisher silo. |
| Litigation analytics on judges and motions | Extensive. A genuine product advantage. | Not offered. |
Two rows are Westlaw wins Marcella has no answer for. Leaving them out would make this page useless to the person it is written for.
Questions
What is the cheapest real alternative to Westlaw?
Free, if your state bar provides a Fastcase or vLex membership tier. More than 40 state bars do. The free tier generally excludes advanced AI features and sometimes the full citator, so confirm what your package includes. Below that, published paid rates start near $99 per seat per month.
Does Marcella replace KeyCite?
Not on editorial depth. KeyCite carries four decades of human classification that no competitor reproduces quickly. Marcella's citation treatment engine maps relationships from the public record across publishers and returns the citing passage, so the attorney reads the later decision rather than a verdict symbol. Those are different products solving overlapping problems.
Can I use Marcella and keep Westlaw?
Yes, and for litigation teams that KeyCite daily that is the sensible configuration. Marcella runs inside Microsoft 365 and does not require removing anything.
Why does Westlaw not publish enterprise pricing?
Firms of eleven or more attorneys receive a quote. Seat count, jurisdictions, content packages, and term all move the number, and multi-year terms carry reported discounts of twelve percent at two years and eighteen percent at three. A published rate lets a buyer compare and walk away. A quote establishes the buyer's budget first.
Where does client data go?
Marcella runs inside the firm's existing Microsoft 365 tenant. On the BOYA tiers it runs in the firm's own Azure tenant. There is no external AI vendor in the data path, Marcella does not train on firm queries, and PII is stripped by Presidio anonymization.
Launch pricing. Rates increase after the launch period. Waitlist members get 30 days free on any paid plan.
Join the Waitlist