Product preview
Name
Systematise
Automate
Illustrative interface, rebuilt for this page - not a live screenshot.
What this is
Every case study on this site runs, in part, on MatterOS - adnah's own matter-management system. You drop in case files and it assembles the matter: parties, chronology, findings, deadlines, first-draft documents, every extracted fact cited back to the page it came from. The lawyer reviews and approves. MatterOS doesn't originate; it proposes, with a citation attached to every claim.
Built by Raghav R Handa, founder of adnah - the agentic operating system every adnah transformation runs on, and the same product he sells directly, independent of the studio. MatterOS on thematteros.com ↗
The challenge
Off-the-shelf practice management software assumes a fixed workflow per practice area and treats AI as a bolt-on feature, not the operating layer. Firms adopting AI-Native transformations needed a matter container flexible enough to hold litigation, corporate, IP, family, and immigration work on one core data model, with agents that could read, extract, draft, and cite - without asking every firm to build that infrastructure themselves.
How we built it
01
One core data model, many practice-area packs
Matter, Actions, Time, Evidence, Research, and Team are the six primitives every practice area maps onto. A litigation matter and an incorporation matter share the same underlying container; only the vocabulary layer on top - the 'practice-area pack' - changes.
02
Extraction with citations, not summaries
When a matter opens, the system reads every uploaded document and extracts parties, dates, and findings - each one linked back to the exact source page. Nothing is asserted without an anchor a reviewer can click through to.
03
Multi-tenant by design, not retrofitted
Row-level security scopes every firm's matters, documents, and financials from every other firm's, from day one - not added later once the product had traction.
04
Credit-metered AI, visible to the firm running it
Every extraction, draft, and search draws from a per-firm AI budget the firm can see and manage, so cost stays predictable as usage scales.
What it does
Smart Inbox
Forward a client email in, and MatterOS classifies it, opens or updates the matching matter, and pre-populates the fields a lawyer would otherwise chase manually.
AI drafter and discovery assistant
Produces first-pass drafts against the firm's own playbooks and surfaces relevant prior documents during discovery, always with source citations attached.
Deadline and clause extraction
Reads incoming documents for dates and obligations and pushes them onto the matter's structured calendar automatically.
Client portal and interview tool
A structured, branded client-facing surface for intake and status updates, so clients get a form and a dashboard instead of a phone tag.
Admin, dashboards, and AI economics
Firm- and user-level admin, usage dashboards, and a live compliance panel mapping the platform to roughly 35 frameworks - SOC 2, ISO, GDPR, HIPAA, and ABA Model Rule 1.6 among them.
MCP server
Exposes matter data through the Model Context Protocol, so a firm's own AI tooling can query MatterOS directly rather than working around it.
Tech stack
- TanStack Start (React 19)
- File-based routing and server functions for the full application, front end and back end in one framework.
- Supabase
- Postgres with row-level security for multi-tenant data isolation, plus auth, storage, and pgvector for semantic search over firm documents.
- AI gateway with per-firm budgets
- Routes extraction, drafting, OCR, and transcription requests through a metered gateway so cost and usage stay visible per firm.
- Framer Motion + Tailwind v4
- The interaction layer - the practice-area packs and dashboards share one design system that flexes per vertical.
Outcome
- Practice-area packs shipped
- 6
- Compliance frameworks mapped
- ~35
- Firms running matters through it
- Every AI-Native case study on this site