Lawkitt Logo

Legal Agents Are Here.
Right Inside Your Editor.

Your team of legal agents inside Lawkitt Editor and integrated to the web, and Slack for unified collaboration.

Download for macOSmacOS installer
Download for WindowsWindows installer

Trusted by lawyers at

Apple Logo
Netflix Logo
Microsoft Logo
Amazon Logo
ByteDance Logo
Rakuten Logo
Carvana Logo

To trust an agent, you have to do it on your own terms.

Lawkitt is designed to give lawyers the confidence to do more with AI.

Model-agnostic by design

Flexible and future-proof.

Legal work varies by matter, risk, and jurisdiction. The best model for the job changes fast.

Lawkitt works with a broad set of providers and local models, so you can choose the right balance of accuracy, speed, and privacy.

Compatible with dozens of providers
OpenRouter LogoAnthropic LogoOpenAI LogoGemini LogoGrok LogoBedrock LogoMoonshot LogoQwen LogoKimi LogoMistral LogoOllama Logo

Role-specific Modes

On-task and under control.

Strong agents need clear roles. Lawkitt's Modes keep models focused on the task at hand and limit tools to what that role needs.

Modes ask to switch when a request moves outside their responsibilities, so your work stays aligned and auditable.

Some examples

  • Orchestrator

    Coordinates multi-step legal workflows via specialized modes

  • Plan

    Creates clear, executable task plans for legal work

  • Edit

    Applies precise edits to documents based on provided instructions

  • Ask

    Answers legal questions without modifying documents

  • File Organizer

    Organizes files and folders for legal workspaces (file operations only)

  • Review

    Review legal document changes with risk-focused feedback

Highly configurable

Make it fit your workflow.

Define guardrails, tool access, and approval steps per matter, client, or team.

Lawkitt is flexible enough to match your review standards, playbooks, and governance requirements.

Secure and transparent

Governed by design.

Keep sensitive matters on your machine when you need to, and control what agents can access at every step.

With clear approvals and traceable actions, you stay accountable and in control.

Two ways to work with Lawkitt.

Use Lawkitt in your editor for hands-on work, or delegate tasks from the web and Slack.

Lawkitt Editor

For focused, hands-on work

Draft, review, and research inside the Lawkitt Editor with built-in agents, full matter context, and clear approval checkpoints:

  • Review and approve every action
  • Maintain matter-aware context
  • Configure workflows and guardrails
  • Preview edits before they land
  • Keep sensitive work on your machine

Ideal for drafting, clause revisions, and fast-turn research.

Lawkitt Cloud + Slack

Coming soon
For collaborative legal workflows

Delegate work to Lawkitt agents from the web or Slack, then review results with full audit trails:

  • Assign research, drafting, and review tasks
  • Collaborate across matters and teams
  • Centralized policies and approvals
  • Track provenance and changes

Ideal for team workflows and multi-stakeholder review.

Request a demo

The AI team that helps your entire human team

Associates, partners, compliance, and ops teams all move faster with Lawkitt's approval-first agents and shared context.

Associate
To Reviewer Agent
Summarize key indemnity changes between these two drafts.
via Matter Files
Paralegal
To Planner Agent
Extract key dates and obligations into a timeline.
via Matter Files
Litigation Counsel
To Researcher Agent
Find recent cases on this jurisdictional issue and summarize holdings.
via Slack
Compliance
To Drafter Agent
Draft a policy update for the new regulation and flag gaps.
via Web
In-House Counsel
To Reviewer Agent
Review this MSA and flag non-standard risk terms.
via Editor
Contracts Manager
To Drafter Agent
Draft a fallback clause for limitation of liability.
via Editor
Legal Ops
To Planner Agent
Standardize our clause library for this template set.
via Editor
Partner
To Planner Agent
Outline negotiation positions for the data privacy addendum.
via Matter Files

Frequently Asked Questions

Everything you need to know about Lawkitt for legal work—privacy, defensibility, and faster drafting.

Lawkitt is an AI system that lives in your editor and matter workspace. It plans work, delegates to specialist modes, drafts/redlines, researches with citations, and reconciles outputs—always with human-in-the-loop approval.
Orchestrator + specialist modes, project-wide indexing, MCP connectors to your knowledge bases, guarded web search, permission-gated actions, and exportable audit logs. Built for defensibility, not one-off chats.
Yes. Lawkitt runs locally by default. No files leave your machine unless you enable a provider. You control redaction and ignore lists, or run local models for maximum privacy.
Model-agnostic. Use OpenAI, Anthropic Claude or Gemini. For enterprise clients, Lawkitt provides local/on-prem LLMs (air-gapped or VPC).
Yes. Core modes (Research, Drafting, Review) pair custom modes (Corporate, Employment, IP, Contract, Privacy, Litigation and more). Connect company playbooks, clause banks, and templates.
Just install Lawkitt and sign in.
Yes. Lawkitt uses efficient context strategies (summarization, targeted retrieval) and links analyses back to source documents for verification. It scales to large document sets.
Designed for compliance workflows: bring‑your‑own model (including on‑prem), permission‑gated actions, and exportable audit logs of prompts, outputs, and approvals.
Yes. Agents can execute multi‑step plans. When information is missing or a decision is needed, they ask a follow‑up question and pause for your input (human‑in‑the‑loop).
The Orchestrator delegates tasks to the best mode and agents can request a role switch when it improves outcomes. You can oversee or approve switches as needed.
Purpose‑built tools to read files, edit files, and run search‑and‑replace; project indexing and semantic code search; connectors to MCP servers for internal/external knowledge; and optional web search.
All popular text formats: Markdown/MDX, plain text, DOCX, XLSX, PDF, CSV/TSV. Agents operate directly on your files with precise diffs.
Yes. Your workspace is indexed for fast retrieval so agents can cite and navigate real files. Private by default—only shared externally when you explicitly allow it.
Yes. Through MCP connectors, agents can query internal wikis, databases, or APIs and merge those results with local project context — always under your control.
When enabled, agents can search the web and bring back sources with citations, combining external findings with your code and documents.
Human-in-the-loop by design: agents ask targeted follow-ups for critical decisions and pause at approval checkpoints. Every action is logged with inputs, sources, outputs, and approvals. Checkpoints let you safely apply edits and roll back. Use History to review past tasks and resume where you left off. All changes are transparent and auditable.

Move faster with Lawkitt.