What it is
Kira Systems is the contract-analysis platform that pioneered ML-based clause extraction for M&A due diligence and large-scale contract review. Acquired by Litera in 2021, it now lives inside the Litera platform alongside Litera Transact (deal management) and Litera Compare (document comparison). In January 2026 Litera shipped a hybrid model: Kira’s proprietary clause models (trained on a stated 45,000+ lawyer-hours) now run alongside a generative-AI layer. Deployed across most of the AmLaw 100 for transactional and diligence work.
Why it shows up in Legal Ops stacks
- Clause extraction depth, now with generative fields. Hundreds of pre-trained clause models (change of control, assignment, indemnification, governing law, exclusivity) plus custom models trained on a firm’s own corpus. The January 2026 Generative Smart Fields let you create a new extraction field from a natural-language description instead of labeling training documents — the setup that used to take days drops to minutes.
- Built for diligence at volume. Pull 5,000 contracts into a data room, run Kira, get a structured matrix of every key clause across every contract. Concept Search and Grid Chat let a reviewer query the review set in natural language. Litera cites up to 50% time savings on review (vendor figure) — the point is replacing weeks of associate red-flag time.
- Litera ecosystem fit, plus bundled Lito. Pairs with Litera Transact for deal workspaces and Litera Compare for redline workflows — one vendor relationship for the transactional stack. Kira now also bundles Lito, Litera’s agentic legal AI, at no extra cost: Lito handles fast everyday reviews in Word, Outlook, and the web, while Kira does high-precision analysis on higher-stakes matters.
Pricing
- Custom quote only — no public price. Sold per-attorney/year and usually packaged with other Litera products; standalone Kira is priced less favorably than the bundle. Third-party estimates (not vendor-published) put small teams and single diligence engagements well below firm-wide deals, while AmLaw and Magic Circle platform commitments reach high-six to low-seven figures annually.
- Lito is included at no additional cost as of mid-2026, which shifts the math when you compare Kira’s bundle against a standalone LLM-review tool priced per seat.
- Implementation is typically 60-90 days with Litera professional services; Generative Smart Fields cut the custom-field labeling that used to drive much of that timeline.
Best for
- M&A and transactional practices doing high-volume diligence
- Litigation teams handling document-heavy second requests and HSR responses
- In-house legal teams reviewing inherited contract portfolios after acquisitions
- Firms already standardized on Litera Transact and Compare that want a single contract-intelligence vendor
Alternatives and watch-outs
- Pick a different tool if extraction isn’t your center of gravity. For LLM-native review, drafting, and negotiation rather than extraction-first diligence, Harvey (broad legal-AI suite) or Luminance (AI-native diligence and review) fit better. Hebbia is the faster-growing entrant to weigh for structured extraction over deal-room corpora when you aren’t tied to the Litera platform.
- Opaque, bundle-driven pricing. The effective cost after add-ons and renewals runs well above the headline per-seat rate. Guard: get standalone Kira and full-platform quotes in writing, and model the three-year cost including renewal uplift before signing.
- Lito and Kira “operate as separate tools today.” The agentic layer isn’t one surface with Kira’s extraction grid yet. Guard: map which work goes to Lito (everyday Word/Outlook review) versus Kira (high-stakes diligence) before rollout, so reviewers aren’t guessing.
- The 90% accuracy figure is a vendor claim, not an independent benchmark, and generative extraction still misreads edge-case clauses. Guard: keep a human check on extracted values for high-stakes provisions, and run your own labeled test set before trusting a new Smart Field in production.