Comparisons
Spellbook vs. Harvey AI: Contract Drafting Tool or Big Law Platform?
Spellbook builds AI contract drafting inside Microsoft Word for transactional lawyers and in-house teams. Harvey sells a broader generative-AI platform for research, drafting, and due diligence to the world's largest law firms. Here is how the two actually differ, in product, customer, and scale.
Spellbook and Harvey are both well-funded legal AI vendors, but they are not really built for the same buyer. Spellbook is a Microsoft Word add-in focused on contract drafting and review, sold seat-by-seat to transactional lawyers and in-house teams. Harvey is a broader generative-AI platform spanning research, drafting, bulk document review, and workflow automation, sold top-down to large law firms and backed from its earliest days by the OpenAI Startup Fund.
Legal AI coverage tends to lump every vendor into one "AI for lawyers" category, which flattens a real strategic split in the market. Spellbook CEO Scott Stevenson has described that split directly: broad, chat-shaped platforms like Harvey sell top-down into large law firms, while workflow-embedded tools like Spellbook go bottom-up through individual lawyers as a plugin inside software they already use every day. Understanding that split, more than any single feature checklist, is the fastest way to figure out which platform actually fits a given legal team.
What is Spellbook and how does it work inside Microsoft Word?
Spellbook began life in Toronto in 2018 as a company called Rally Legal, then rebranded to Spellbook in 2022 alongside the launch of a GPT-3-powered Word add-in, among the first generative-AI tools built specifically for lawyers. Co-founder and CEO Scott Stevenson has compared the product to GitHub Copilot for coders, or an "electric bicycle" for legal drafting: a lawyer still steers and pedals, but the AI makes the uphill parts of contract work faster. The platform's core capabilities sit directly inside a Word document: Review flags redlines, aggressive terms, and missing clauses; Draft generates new clauses or full sections from templates or prior precedent; Ask answers questions about a contract with citations linked back to source language; and Spellbook Associate, a newer agentic layer, handles more autonomous first-pass drafting akin to a junior associate. Spellbook reports more than 4,000 customers across more than 80 countries and says it has reviewed over 10 million contracts since launch, with in-house legal teams — including corporate clients such as Nestlé — now generating more than half of its revenue, a shift from its original law-firm-only customer base.
What is Harvey AI and how does it serve large law firms?
Harvey was founded in 2022 by Winston Weinberg, a former litigator at O'Melveny & Myers, and Gabriel Pereyra, a former research scientist at Google DeepMind and Meta. The OpenAI Startup Fund became Harvey's first institutional investor, anchoring a seed round shortly after the founders met with OpenAI's leadership, and Harvey was given early access to GPT-4 ahead of its public launch. Allen & Overy (now A&O Shearman) was Harvey's landmark early customer, running a trial in which roughly 3,500 lawyers logged around 40,000 queries; PwC later adopted Harvey for thousands of legal professionals operating across dozens of countries. Harvey's product suite is built around four pieces: Assistant for chat-based drafting and analysis, Vault for bulk, cross-document review used heavily in M&A due diligence, Knowledge for legal and regulatory research with citations, and Workflow Agents for building multi-step automated processes without code, plus integrations into Word, Outlook, SharePoint, and Box. The company holds SOC 2 Type II and ISO 27001 certifications and offers data residency in the United States, European Union, Switzerland, and Australia, reflecting how much of its buyer base sits inside large, compliance-sensitive law firms and enterprises.
How do Spellbook and Harvey differ in target customer and scope?
| Dimension | Spellbook | Harvey |
|---|---|---|
| Founded | 2018 (as Rally Legal; relaunched as Spellbook 2022), Toronto | 2022, San Francisco |
| Core interface | Microsoft Word add-in | Standalone platform plus Word/Outlook/SharePoint integrations |
| Primary buyer | Individual transactional lawyers, in-house legal teams | Enterprise law firms, Am Law 100, corporate legal departments |
| Core workflow | Contract drafting, review, redlining, clause benchmarking | Research, drafting, bulk due diligence, workflow automation |
| Adoption pattern | Bottom-up, per-seat | Top-down, enterprise rollout |
| Notable customers | Nestlé and other in-house legal teams; commercial law firms | Allen & Overy / A&O Shearman, PwC, Ashurst |
How do funding, scale, and valuation compare?
The two companies are not remotely the same size financially. Spellbook raised a $50 million Series B in October 2025, led by Khosla Ventures, valuing the company at roughly $350 million and bringing its total funding to more than $120 million across four rounds, plus a separate $40 million debt facility from RBCx earmarked for acquisitions. Harvey operates on an entirely different scale: it raised a $300 million Series E in June 2025 at a $5 billion valuation co-led by Kleiner Perkins and Coatue, was confirmed at an $8 billion valuation by December 2025, and then closed a further $200 million growth round in March 2026 at an $11 billion valuation, co-led by GIC and Sequoia Capital, taking total funding past $1.2 billion. Harvey's own announcement and independent analysis both point to well over 1,000 customers across dozens of countries and a majority share of the Am Law 100 as clients. Spellbook's growth has been real but is happening at a materially smaller scale — a distinction that matters because it tends to track with how much platform breadth, security certification depth, and dedicated customer-success support a buyer can expect at each price point.
Which platform should a law firm or legal team actually choose?
The honest answer is that the choice depends on which workflow is actually the bottleneck. A commercial team drowning in NDAs, vendor agreements, or lease reviews will likely see faster time-to-value from Spellbook, because it requires no new login or ecosystem — lawyers keep working in Word while the AI drafts and flags issues around them. A large firm or in-house department whose pain point is research, litigation support, or due diligence across thousands of documents in an M&A data room is closer to Harvey's core use case, and its Vault and Workflow Agents modules are purpose-built for exactly that kind of bulk analysis. Some legal departments run both: Spellbook for high-volume day-to-day drafting, Harvey for research and diligence at scale. Given that neither company publishes list pricing, the practical next step for either platform is the same — request a demo scoped to a firm's actual contract volume, document sets, and practice areas, rather than comparing marketing pages alone.
Frequently asked
What is the main difference between Spellbook and Harvey AI?
Spellbook is an AI add-in that lives inside Microsoft Word, built specifically for contract drafting, review, and redlining by transactional lawyers and in-house counsel. Harvey is a broader generative-AI platform, with separate modules for chat-based research and drafting, bulk document review across thousands of files, legal research with citations, and no-code workflow automation. Spellbook sells seat-by-seat to individual lawyers and legal departments doing high-volume contract work. Harvey sells top-down enterprise deployments to large law firms and corporate legal departments that need research, due diligence, and multi-practice coverage, not just contracts. They compete for some of the same budget but were built for different daily workflows.
Which law firms and companies use Harvey AI?
Harvey's landmark early customer was Allen & Overy (now A&O Shearman), where roughly 3,500 lawyers ran about 40,000 queries during an early trial. PwC later adopted Harvey for thousands of legal professionals across dozens of countries, and firms including Ashurst and A&L Goodbody have also partnered with the platform. By 2025 and into 2026, Harvey reported serving well over a thousand customers across dozens of countries, including a large share of the Am Law 100 and major professional-services firms. Its customer base skews toward large law firms, in-house legal departments at multinational companies, and financial institutions that need research and due-diligence capacity at scale.
Who uses Spellbook, and is it only for law firms?
Spellbook started by selling to law firms doing transactional and commercial work, but its customer mix has shifted: in-house legal teams now generate more than half of Spellbook's revenue, with corporate clients including Nestlé. The company reports thousands of customers across dozens of countries, with the United States making up the majority of usage and the United Kingdom growing quickly. Because Spellbook works inside a document lawyers already have open, it is commonly adopted bottom-up by individual attorneys or small legal teams handling contract review, rather than rolled out top-down as a firm-wide enterprise platform the way Harvey typically is.
How much does Spellbook cost compared to Harvey AI?
Neither company publishes standard per-seat pricing; both quote custom terms based on team size and features. Spellbook sells seat-based subscriptions, with pricing scaling as more team members are added to a license, and independent legal-tech pricing trackers describe multiple tiers ranging from individual plans up to enterprise tiers with seat minimums. Harvey similarly negotiates enterprise contracts directly with law firms and corporate legal departments, scaled to lawyer count and module access such as Vault or Workflow Agents. Prospective buyers of either platform should expect a sales conversation and a demo rather than a published price list, and should request a quote scoped to their actual usage.
Is Harvey AI or Spellbook backed by OpenAI?
Harvey has the direct OpenAI relationship: the OpenAI Startup Fund was its first institutional investor, anchoring a seed round shortly after Harvey's founders met with OpenAI's leadership in mid-2022, and Harvey received early access to GPT-4 ahead of its public release. Spellbook uses large language models under the hood as well, but it is not an OpenAI-backed company; its investors instead include Khosla Ventures, which led its 2025 Series B, along with Thomson Reuters Ventures, Inovia Capital, and other legal-tech and venture funds. The OpenAI relationship is a meaningful part of Harvey's origin story and fundraising narrative in a way that has no direct parallel at Spellbook.
Can a law firm realistically use both Spellbook and Harvey?
Yes, and some legal departments effectively do, because the two tools are not perfect substitutes. A transactional team might keep Spellbook open in Word for day-to-day contract drafting, redlining, and clause benchmarking, while a firm's broader research, litigation-support, or M&A due-diligence teams use Harvey's Vault and Workflow Agents to process large document sets or draft research memos with citations. Budget-constrained legal departments often start with one tool matched to their most painful workflow, then evaluate the other later. The right starting point depends on whether the bigger cost center is contract volume or research and diligence workload.