Skip to main content

Competitive analysis

Kervalt vs Libra

Libra is a legal AI workspace for Dutch law firms, 1,300+ teams, with a chat-first assistant, document review, and citation-on-every-answer grounding. Developed by lawyers, for lawyers. Here's how it stacks up against Kervalt, honestly.

Investigation date: October 22, 2026 · Source: libratech.ai marketing and in-app navigation

Kervalt wins

7

Parity or close

3

Gaps to close

5

Capability × 15

Capability Libra Kervalt Verdict
Chat with research/tools modes Core product prompt / loop / graph: stronger Kervalt ahead
Citations on every answer Citation chips on every answer Architecture in place, not yet visible in the UI Close race
Review mode (document redline/diff) Document review built in Not yet Gap
Discovery (legal research surface) Legal discovery across Dutch sources Source retrieval only Gap
Templates gallery Built-in gallery Library Parity
Word add-in / DOCX round-trip Works inside Word Markdown / HTML only Gap
Legal data connectors (case law, legislation, KvK) Implied by the product Not yet Gap
⌘K project search Built in Built in Parity
Case / matter organization Projects Cases + board + scaffold Kervalt ahead
Human-in-the-loop gates Chat + citations only Core: Approve / Send / Sign gates Kervalt moat
Agent scoping & SOP state machine — Core Kervalt moat
Capability map as operating spine — Business Capability Map 2.0 Kervalt moat
Multi-industry (law, trading, recruiting) Legal only Personas per industry Kervalt ahead
Self-hosted / on-prem single binary SaaS only Rust binary with embedded assets Kervalt moat
EU / GDPR compliance packaging Marketed on their site Audit in place; DPA / ISO story coming Gap

The strategy

Don't clone the chatbot. Flank it

Libra is a better chat window; Kervalt is an operating system. We compete by closing the visible UX gaps law firms expect, and by positioning on what Libra structurally cannot do: gates, scoped agents, capability economics, and running on your own server.

  • Close the visible gaps: review mode, visible citations, legal source connectors, DOCX export.
  • Position on the moats: human gates, scoped agents, the capability map, on-prem deployment.

Gap backlog

What we're closing, and in what order

P0: must-have

To be shoppable by law firms

  • Review / Diff mode

    Side-by-side original vs AI proposal, per-change accept/reject, tracked-change export. The single most-demanded legal AI feature.

  • Visible citations in every AI answer

    Clickable citation chip that deep-links to the source page and chunk. Makes the anti-hallucination architecture marketable, not just real.

  • Legal source connectors

    New source types: legislation, case law (ECLI), and register data (KvK). Start with demo groups; production gets API harvesters with version stamps.

  • DOCX export

    Every Markdown deliverable exports to .docx. Lawyers live in Word; a deliverable they can't send is a deliverable that doesn't exist.

P1: this quarter

Competitive parity

  • Discovery surface

    A dedicated research view: ask once, see retrieval ranked across sources, save the result as a file into a capability.

  • Word add-in

    Long-term: a real add-in. Interim: copy-as-tracked-changes paste.

  • Compliance pack

    DPA generator, data-residency statement, ISO 27001 roadmap, zero-retention LLM config, audit-log export. The self-host binary becomes the reason enterprises pick Kervalt.

  • Dutch UI + legal Dutch prompt tuning

    Libra is Dutch-first; we meet customers in their language.

P2: amplifiers

Differentiation

  • Per-case conversation threads

    Scope the AI panel to a case, like Libra's chat history, but organized around the matter, not the chat.

  • Templates gallery UX

    A proper gallery experience on top of the Library.

  • Academy / Getting Started

    Package the onboarding scenario as a guided academy.

Positioning

What to say against Libra

“Libra answers. Kervalt operates. Agents run your capability map and stop at your gates.”

Kervalt

“Their AI talks to you. Ours works for you: scoped, supervised, attributable.”

Kervalt

“SaaS in their cloud, or your binary on your server? Legal privilege likes the second one.”

Kervalt

Want the full investigation, or a comparison against another tool? Talk to us.

NEXT STEP

Ready to see the full picture?

See how your team finds hidden risks and decides with confidence, in minutes.