Spellbook

AI contract drafting and review inside Microsoft Word — the SMB / mid-market answer to Harvey, distributed via the Word add-in surface.

1. Core Product / Service

Spellbook is an MS Word add-in (formerly "Rally Legal") that surfaces clause suggestions, redlines, risk flags, and term extraction directly inside the lawyer's drafting workflow [1][2]. It runs on a frontier-model stack (GPT, Claude) routed by task. Core capabilities:

• Clause suggestion while drafting — autocomplete legal language with playbook-aware context.
• Risk flagging on uploaded counterparty drafts — highlight unusual / unfavorable terms vs. the firm's playbook.
• Redline generation — propose markup compatible with Word's track-changes model.
• Benchmarking — Spellbook references a curated corpus of millions of contract clauses to flag outliers.

The Word-native distribution is the deliberate moat: lawyers don't switch tools, they get AI in the place they already draft. Contrast with Harvey (separate web app) and Robin (separate platform).

2. Target Users & Pain Points

• Buyer: Solos / boutiques / SMB-to-mid-market law firms; in-house counsel at startups and mid-market companies who draft & negotiate contracts in Word.
• Pain points addressed: Junior associates spend hours redlining vendor contracts and NDAs; Spellbook compresses that to minutes without forcing the lawyer out of Word.
• Wedge: Distribution. Word is the universal legal drafting surface; Spellbook plugs into existing habit rather than asking lawyers to switch to a SaaS web app.

3. Competitive Landscape

Spellbook's structural advantage: cost-per-seat is ~10× lower than Harvey while addressing 80% of common drafting needs for SMB users.

4. Unique Observations

• Pricing model: Per-seat SaaS, three tiers — Starter $99/mo, Professional $149/mo, Enterprise $199/mo (10-seat minimum). All annual; monthly billing adds ~20%. No free tier; 14-day trial [2].
• Implied $/1M tokens consumed: Drafting users likely consume ~1-3M tokens/seat/month (clause completions are short, but cumulative across full days of drafting). At $149/mo, retail rate is roughly $50-$150 per 1M tokens. Frontier API cost (2026): ~$3-$15/1M. Markup ~10-30×. Much lower retail unit price than Harvey, reflecting the SMB segment's price sensitivity.
• Moat type: Distribution (Word add-in surface) + workflow integration. Not a model moat. The Word footprint is a real wedge against Harvey, but vulnerable to Microsoft Copilot for legal bundling.
• Customer profile: SMB to mid-market lawyers; transactional practice (M&A, commercial, real estate) where drafting volume is high and seat budgets are constrained. Explicitly not BigLaw — Harvey owns that segment and Spellbook doesn't try to compete on enterprise procurement.
• Markup multiple over raw API: ~10-30×. Sustainable because the buyer compares to associate-hour cost ($150-$400 billable for SMB associates), not to API cost.
• Funding context: Total $84.8M over 4 rounds; $40M Series B in Mar 2026 [5]. Materially less capital than Harvey ($800M+) but the segment requires less GTM spend — SMB closes faster, lower ACVs, lower CAC.

5. Financials / Funding

• Total raised: ~$84.8M cumulative across 4 rounds [5][6].
• Latest round: $40M (Mar 2026); prior $20M+ raises through 2022-2024.
• Investors: Inovia Capital (lead); Bain Capital Ventures; Thomson Reuters Ventures; Moxxie Ventures.
• Founded: 2018 (originally as Rally Legal); rebranded to Spellbook around the AI launch (~2022).
• Customer count: Not disclosed publicly; positioned as the leading Word-native legal AI add-in.

6. People & Relationships

• Founder: Scott Stevenson (CEO) — Canadian (St. John's, Newfoundland-based originally), legal-tech founder; vocal advocate for SMB-tier legal AI.
• Investors: Inovia, Bain Capital Ventures, Thomson Reuters Ventures, Moxxie.
• Channel partners: Microsoft (via Word add-in); Thomson Reuters (via investor + ecosystem).
• Related wiki: harvey (BigLaw counterpart), robin-ai (mid-market peer that didn't survive 2025).