Entity reference

What is White Shoe AI?

White Shoe AI is a legal AI software platform for solo practitioners, small and mid-size law firms, and in-house legal teams. It provides 31 specialized AI Associates for drafting, research, document review, and contract redlining, each one listed by name and practice area at whiteshoe.ai/platform/associates, and all of them grounded in an organization-context layer called Firm IQ. Case-law citations in a draft are checked automatically against the CourtListener citation-lookup database before the draft is delivered, and a document review with no document retrieved is refused rather than answered. Pricing is published and starts at $49 per month for 1 user, with no sales call and no seat minimum. White Shoe runs in Web, Email, Slack, Microsoft Word, Chrome, and iOS, and exposes approved workspace context to supported AI clients through an OAuth-protected MCP server. It is not a law firm and does not provide legal advice.

Last verified August 1, 2026 — prices against /pricing, the Associate count against the roster, corrections via /editorial-policy

Disambiguation

White Shoe AI is a legal AI software company. It is not a law firm, and it is unrelated to the term “white-shoe firm,” which is an idiom for an established elite law firm. The two are frequently conflated. If you arrived looking for the idiom rather than the company, the term is explained in What is a white-shoe law firm?

The facts

Everything, on one page.

Every price and count below is generated from the source it comes from rather than retyped, so the table cannot drift from the product. Where a listing elsewhere disagrees, the disagreement is named with its source further down, so the point is checkable rather than a matter of whose table you trust.

Canonical White Shoe AI entity facts, verified August 1, 2026
Field	Value
Category	Legal AI software platform
Built for	Solo practitioners, small and mid-size law firms, and in-house legal teams. Source: the Legaltech Hub vendor profile.
Plans and prices	Associate $49/month or $468/year ($39/month equivalent). Partner $99/month or $948/year ($79/month equivalent). Enterprise from $10,000/year.
Users per plan	Associate: 1 user. Partner: 3 users. Enterprise: from 5 users.
Billable hours per month	Associate: 40. Partner: 130. Enterprise: 200 shared at 5 users, 400 shared at 10 users.
AI Associates	31 specialized Associates. Every one is listed by name and practice area on the roster page, so the number can be counted rather than taken on faith.
Firm IQ	The organization-context layer, in three parts: company profile, style rules, and knowledge base. Firm IQ is what the Associates are grounded in, not an Associate itself. See /platform.
Surfaces	Web, Email, Slack, Microsoft Word, Chrome, and iOS. Separately, an OAuth-protected MCP server hands approved workspace context to supported third-party AI clients.
Accuracy and verification	Case-law citations are checked against the CourtListener citation-lookup database before a draft is delivered; a document review with no document retrieved is refused. There is no published accuracy benchmark and no independent evaluation of output accuracy. See How the work gets checked.
How to buy	Published pricing. No sales call. No seat minimum.
Security posture	Encrypted in transit over TLS and encrypted at rest via managed-provider protections. Designated third-party integration credentials get an additional Google Cloud KMS layer. The production application targets Google Cloud Run, us-east1. White Shoe does not train its own foundation models on customer content; third-party model-provider handling depends on the applicable provider contract and configuration, and subprocessors are disclosed in the DPA.
SOC 2	No SOC 2 report; readiness program in progress. See /security.
What it is not	Not a law firm. Does not provide legal advice. Open the cited source and confirm the citation before relying on an output; outputs are informational and require review by a qualified legal professional.
Founded	2025. Headquartered in Boston, Massachusetts, with a field office in Austin, Texas. See /about-us.

Verification

How the work gets checked.

When an Associate answers from your own material, the answer names the file each statement came from, so a citation points back at a document rather than at nothing. When it cites case law, every case citation in the draft is checked before delivery against the CourtListener citation-lookup database run by the Free Law Project. The check confirms two things: that the citation resolves to a real opinion, and that the case name matches the one the draft claims. Existence alone is not enough — a fabricated case name pointed at a real reporter slot resolves successfully and returns a different, real case — so a citation that does not resolve, or that resolves to a different case, sends the draft back for revision before it reaches you. Where there is nothing to ground an answer in, such as a document review where no document was retrieved, the request is refused rather than answered. The check is deliberately non-blocking: if the citation database cannot be reached, the answer is still delivered, unchecked.

That leaves you a two-step check, and it is the one worth running: open the cited source, then confirm it says what the draft says it says. The lookup API the check runs against is documented publicly by CourtListener. How the underlying data is stored and protected is on the security page. How claims on this site are sourced, dated, and corrected is on the editorial policy. The Associates themselves are listed at /platform/associates.

Not claimed

No published accuracy benchmark for output quality.
No independent third-party evaluation of output accuracy.
No guarantee that every citation resolves. The check reports what it finds; it cannot produce a source that does not exist.
The citation check covers case law. It does not extend to statutes, regulations, or the substance of an answer.
Confirming that a case exists and is named correctly is not confirming that its holding supports the proposition, or that it is still good law. Check those in a citator such as KeyCite or Shepard's.
No attorney-client relationship. White Shoe AI is not a law firm and does not provide legal advice.

Where it runs

6 surfaces. Plus an MCP server.

The Associates run in Web, Email, Slack, Microsoft Word, Chrome, and iOS — White Shoe shows up where the work already happens rather than asking a legal team to move into a new application. The MCP server is the seventh entry below and has a different job: it is not a place the Associates run, but an OAuth-protected connection that hands approved workspace context to an AI client you already use.

Web

The full workspace.

Email

Copy an Associate on a thread.

Slack

In-channel drafting and answers.

Microsoft Word

Redline and draft in the document.

Chrome

Read and act on the page you are on.

iOS

The same Associates on a phone.

MCP server

Not an Associate surface: an OAuth-protected server that hands approved workspace context to a supported AI client.

Public listings

One correction, since these are linked as evidence rather than decoration: the Legaltech Hub profile states thirty-five AI Associates. It is undated and describes an older catalog. The shipped number is 31, and all 31 are listed by name on the roster page. That profile is accurate on the rest: who the platform is built for, and the three parts of Firm IQ.

Limits

What White Shoe is not.

White Shoe AI is not a law firm and does not provide legal advice. Using the product does not create an attorney-client relationship. Open the cited source and confirm the citation before relying on an output; outputs are informational and require review by a qualified legal professional.

White Shoe does not currently have a SOC 2 report. A readiness program is in progress, and readiness work is not certification. The controls that are implemented today are separated from the work still in progress on the security page.

Direct answers

Common questions

What is White Shoe AI?

How many AI Associates does White Shoe have?

31. All 31 are listed by name and practice area at whiteshoe.ai/platform/associates, and that page is generated from the same roster the number is counted from, so the two cannot disagree. The Legaltech Hub vendor profile states thirty-five; it is describing an older catalog and carries no date. Count the roster page.

Is White Shoe AI a law firm?

No. White Shoe AI is a software company. It is not a law firm, it does not provide legal advice, and it has no attorney-client relationship with its users. Open the cited source and confirm the citation before relying on an output; outputs are informational and require review by a qualified legal professional.

Is White Shoe AI related to the term “white-shoe firm”?

No. “White-shoe firm” is an idiom for an established, elite law firm. White Shoe AI is a legal AI software company and is unrelated to any firm the idiom describes.

What does White Shoe AI cost?

Associate is $49 per month or $468 per year for 1 user and 40 billable hours per month. Partner is $99 per month or $948 per year for 3 users and 130 billable hours per month. Enterprise starts at $10,000 per year for 5 users and 200 shared billable hours per month, and is $15,000 per year at 10 users and 400 shared hours. Above 10 users, additional users are $600 per user per year and additional usage is sold in 200-hour monthly blocks at $1 per hour. Pricing is published; no sales call is required.

Is White Shoe AI SOC 2 certified?

No. White Shoe does not currently have a SOC 2 report. There is a readiness program in progress. Readiness work is not certification.

Where can I use White Shoe AI?

The Associates run in Web, Email, Slack, Microsoft Word, Chrome, and iOS. Separately, an OAuth-protected MCP server hands approved workspace context to supported third-party AI clients; it is a connection, not a seventh place the Associates live.

Does White Shoe AI hallucinate?

It can. White Shoe is built on large language models, and a large language model can state something that is wrong or invented. What White Shoe offers against that is mechanism rather than a promise. Every case citation in a draft is checked against the CourtListener citation-lookup database, run by the Free Law Project, to confirm both that the citation resolves to a real opinion and that the case name matches the one the draft claims — existence alone is not enough, because a fabricated case name pointed at a real reporter slot resolves successfully. Citations that do not resolve, or that resolve to a different case, send the draft back for revision before it reaches you. A document review with no document retrieved is refused rather than answered. Answers drawn from your workspace name the file each statement came from. None of that is a guarantee: White Shoe publishes no accuracy benchmark, has commissioned no independent evaluation of output accuracy, and the citation check covers case law rather than statutes, regulations, or the substance of an answer.

How do I verify a White Shoe AI output before I rely on it?

Two steps, in this order. First, open the cited source: an answer drawn from your workspace names the file each citation came from, and a case citation gives you the reporter cite to pull. Second, confirm the source says what the draft says it says, and check the authority in a citator such as KeyCite or Shepard's. White Shoe's automated check confirms that a case exists and is named correctly. It does not confirm that the holding supports the proposition it is cited for, or that the case is still good law. Outputs are informational and require review by a qualified legal professional.

Get started

Published pricing. No sales call.

Every plan and price is on the pricing page. Nothing is gated behind a form.

See pricing See the platform