Litera

What it is

Litera is the document-and-drafting infrastructure most large law firms already run — Compare, Draft, Kira, Transact, Foundation, DocXtools, pdfDocs and roughly two dozen other modules, assembled through the 18 acquisitions Tracxn catalogues as of April 2026, under Hg majority ownership since May 2019. On 15 July 2026 the company relaunched around a single pitch: one AI agent, called Lito, and one dataset spanning both the practice and the business of law. CEO Avaneesh Marwaha framed the position against the funded startups directly — “Every AI startup in legal is selling a feature. We’re selling the platform those features run on.”

Read the relaunch for what it is. Litera did not ship a new platform in July; it renamed the entry point to one it already had. The catalog still lists Kira, Litera Compare, Litera Transact, Foundation and the rest as separate SKUs, now grouped under Practice of Law, Business of Law and Corporate Law. Artificial Lawyer called it “more of a marketing and brand awareness shift” than a new Litera, and that reading holds up against the product pages.

What is genuinely new is the licensing. On 27 October 2025 Litera put Lito into Draft Base, Draft Pro, Draft Advanced and Kira at no additional cost to new and existing customers, with a baseline of 11 legal AI skills across capital markets, M&A, business development, litigation and compliance. At ILTACON in August 2026 Lito’s scope was restated as five connected capabilities — Draft, Review, Negotiate, Protect, Close — alongside two new products, Firm AI Search and Business Development Intelligence.

Why it shows up in Legal Ops stacks

• Comparison is the lock-in, and it did not move. Litera reports 10 million document comparisons a month, 99% coverage of the Am Law 100 and 99% logo retention (vendor figures, 15 July 2026), and markets the Draft Base comparison engine as trusted by 75% of the market — also a vendor claim, with no published basis. Whatever the exact share, transactional redline cycles at large firms run through this engine, and that is why Litera survives shortlists it would lose on AI alone.
• Lito runs in the applications lawyers already have open. Word, Outlook, Google Workspace, Chrome, Edge, Safari, iOS, the Litera desktop apps, and directly against iManage and NetDocuments repositories. No second login and no second corpus is the whole argument, and for a firm that has already lost one AI rollout to non-adoption it is the strongest one here.
• The generative layer has an off switch. Litera states that generative AI can be disabled entirely and the rules-based engines keep working, with model routing across OpenAI, Anthropic and Gemini under ISO 27001, SOC 2 Type 2, SOC 3, GDPR, NIS 2 and DORA. For a firm whose partners will not put client documents through a probabilistic model, that is a procurement answer Harvey and Legora cannot give.
• Business-of-law coverage no drafting rival matches. Foundation 365 — the Peppermint Technology acquisition of February 2025 — is deployed at five of the ten largest global firms. CRM, pitch, scoping and relationship intelligence sit on the same dataset as the drafting modules.

Pricing

Quote-only, and verified quote-only on 8 September 2026. litera.com/contact-us routes to a demo form, and litera.com publishes no rate card anywhere.

There is a store — store.litera.com sells Draft Base, Draft Pro, Clean and pdfDocs as SKUs with an add-to-cart flow — but prices render only behind a login, so it is not public pricing. Third-party directories put Litera Compare near $195/user/year and Draft Pro near $1,900; both are aggregator estimates with no stated billing unit, not vendor-confirmed rates, and neither should carry a business case. Firms above roughly ten lawyers are quoted per-seat enterprise subscriptions through sales.

The variable that actually moves the quote is the module count, not the seat rate. Most firms buy three to five modules, and Draft Pro alone now carries Litera Compare Desktop, Contract Companion, DocXtools Companion and Best Authority — which is where a “just Compare” conversation quietly becomes a platform conversation. Price the modules first, then the seats.

Best for

A firm IT or legal ops lead at an Am Law 200 or large mid-market firm that already runs Litera Compare and needs the AI rollout to land where lawyers already work. Lito’s inclusion at no extra cost makes the incumbent option cheaper than any competing pilot, and adoption — not model quality — is what kills most firm AI programs.

Do not buy the platform for the AI. If the deciding requirement is research depth, novel legal reasoning or firm-specific agent workflows, Litera is not the strongest product in the room, and Marwaha’s own pitch concedes that ground. Firms under ten lawyers should buy a Draft package from the store and stop there.

Versus the alternatives

• Harvey — the better pure legal-AI assistant, and it ships an MCP server Litera does not. Pick Harvey when the work is research, drafting from scratch and firm-specific agent workflows.
• Thomson Reuters CoCounsel — pick it when Westlaw is the anchor and research plus review should sit on one bill.
• Legora — the fastest-growing entrant in firm-side legal AI, and the vendor Marwaha names as the threat. Pick it when you want an AI-native review surface and will keep Compare separately.
• Intapp — the head-to-head on the business-of-law side against Foundation 365. Intapp is stronger on conflicts, intake and risk; Litera is stronger once drafting is in scope. Full framing in intapp-vs-litera.
• Spellbook and Definely — the cheaper Word-native picks for small firms and in-house teams that will never buy a platform.
• Kira Systems — Litera’s own diligence module, also sold standalone. Read that entry before agreeing to buy it inside a bundle.

Watch-outs

• The relaunch is brand-level; the consolidation behind it is not finished. Kira and Lito still operate as separate surfaces, and Kira’s clause models are ML-era with a generative layer added on top rather than a rebuild. Guard: during the pilot, run one real matter end to end across Draft, Kira and Transact, count every handoff that requires re-uploading a document, and make the account team commit dates for the ones that fail.
• Lito is free today and commercially undefined tomorrow. Marwaha has said it is “too early to call the commercial aspects of Lito,” and that optional future skills may be priced separately. Guard: write Lito’s inclusion into the subscription term at a fixed renewal price, and get the mechanism for pricing future skills named in the contract rather than the roadmap deck.
• No published price, and a store whose prices sit behind a login. Guard: get standalone and full-platform quotes in writing at the same seat count, and model the three-year cost including renewal uplift before comparing Litera against a per-seat AI vendor.
• No MCP server for outside agents. Lito consumes MCP connectors under central firm governance, which points the wrong way — your own Claude deployment or internal agent cannot query Litera. Guard: if agent-side access matters, scope the Compare Server API work at developer.workshare.net during evaluation instead of waiting for a connector that has not been announced.
• Twenty-plus modules means the demo is not the quote. Guard: make the account team annotate each demo capability with the module and package it belongs to, then price only those.