Skip to content
← Back·2026 · AI

Legali

AI-native legal platform — three surfaces for survivors, everyday people, and legal teams.

Role
Engineering Contributor
Period
Sep 2025 — May 2026
Stack
React, TypeScript, LLM, RAG, Agents, MCP
Link
Visit ↗
  • React
  • TypeScript
  • LLM
  • RAG
  • Agents
  • MCP
legali.ai
01/03
  1. Legali marketing site — Justice, accessible to everyone, with three product personas
    Legali marketing site — Justice, accessible to everyone, with three product personas — mobile
  2. TeamLegali workspace — agentic legal AI workspace for firms and corporate teams
    TeamLegali workspace — agentic legal AI workspace for firms and corporate teams — mobile
  3. Lea — women-focused legal voice AI companion with app store and Google Play distribution
    Lea — women-focused legal voice AI companion with app store and Google Play distribution — mobile
legali.ai — public hub for three audiences

Overview

Legali is a San Francisco-based legaltech AI startup building three differentiated AI surfaces on a shared substrate — each tailored to a distinct audience along the legal-access spectrum:

  • legali.ai — public hub framing the mission and routing visitors to the right product.
  • TeamLegali (team.legali.ai) — agentic legal workspace for law firms, corporate legal teams, and startups: intake, drafting, contract comparison, litigation support, e-signature.
  • Lea (lea.legali.ai) — a women-designed legal voice AI companion that turns lived experiences into rights awareness, records, and legally usable narratives. Shipping on iOS and Android.

My role

Engineering contributor across the surfaces — focusing on the AI substrate (LLM features, RAG over legal documents, agentic workflows) and shared product infrastructure that lets three audience-specific frontends sit on top of the same legal-intelligence layer.

Approach

  • React + TypeScript across the product surfaces; shared component primitives for cross-surface consistency
  • RAG pipelines for legal-document retrieval, drafting assistance, and contract comparison
  • Agentic workflows powering TeamLegali's intake-to-output loops
  • MCP-style tool surfaces that let the agent layer interact with internal data + external legal sources safely
  • Tight collaboration with a small SF team across timezones (Jakarta ↔ SF)

Outcome

Three publicly shipping surfaces, each tailored to a different audience but unified by one AI substrate — a structural answer to the legaltech problem that legal-grade help is usually only available to people who already have lawyers.