Skip to content

lvndry/clausea

Repository files navigation

Clausea

Clausea

Policy document intelligence for SaaS products.
Turn privacy policies, terms of service, and related legal pages into plain-English risk assessments.

clausea.co · Example report · AGPL-3.0

Ask DeepWiki


About

Clausea crawls and analyzes policy documents (privacy policies, terms of service, cookie policies, GDPR notices, DPAs, and more) for SaaS products. It extracts structured signals (data collected, third-party sharing, user rights, retention, consent model) and synthesizes them into a qualitative verdict, and decision-ready summaries.

Live product pages: https://clausea.co/products/{slug} Machine-readable summaries: https://clausea.co/api/products/{slug}/summary.txt

Features

Area What you get
Whole-site coverage Crawls a vendor's full policy surface, not just a single pasted URL
Risk scoring 0–10 score with grade, verdict, and score breakdown
Plain-English summaries Consumer-friendly explainers plus detailed policy overviews
Topic intelligence Stance and evidence across key privacy topics
Compliance signals GDPR and CCPA-relevant clause detection
Source documents Linked, cited policy pages behind every finding
Browser extension Analyze policies from the toolbar (packages/extension)
Developer API FastAPI backend with OpenAPI docs at /docs

Who it's for

  • Individuals checking what an app does with their data before signing up
  • Privacy-conscious users comparing vendor policies
  • Compliance and legal teams auditing third-party SaaS vendors
  • Developers building on the Clausea API or self-hosting the stack

Getting started

Prerequisites

  • Python 3.11+ with uv
  • Node.js 18+ with bun
  • MongoDB (local or Atlas)
  • API keys as needed: Openrouter, Clerk, Pinecone (see env examples)

Setup

git clone https://github.com/lvndry/clausea.git
cd clausea
make setup

make setup installs dependencies, creates env files from examples, and configures pre-commit hooks. Run make help for all commands.

Environment

# Backend — edit with MongoDB URI, LLM keys, Clerk JWKS, etc.
cp packages/backend/.env.example packages/backend/.env

# Frontend — Makefile/setup uses .env.local
cp packages/frontend/.env.example packages/frontend/.env.local

See packages/backend/.env.example and packages/frontend/.env.example for the full variable list.

Run locally

# Frontend + backend
make dev

# Or individually
make run-backend    # http://localhost:8000
make run-frontend   # http://localhost:3000
make dashboard      # Streamlit ops dashboard (http://localhost:8501)
Service URL
Web app http://localhost:3000
API http://localhost:8000
OpenAPI http://localhost:8000/docs
Ops dashboard http://localhost:8501

Extension development

make setup-extension   # once
make extension-dev     # builds to packages/extension/.output/chrome-mv3

Load the unpacked extension from packages/extension/.output/chrome-mv3 in chrome://extensions. See the extension README.

Development

make test      # backend (pytest) + frontend (vitest)
make lint      # ruff + eslint
make format    # ruff format + prettier

Backend-only checks (from packages/backend):

uv run ty check
uv run ruff check
uv run pytest

Contributing: fork, branch, open a PR. See CONTRIBUTING.md and AGENTS.md.

How it works

At a high level, indexing a product runs a three-step pipeline:

  1. Crawl — two-pass discovery (precision-first, recall fallback) to find and persist policy pages
  2. Analyze — evidence-first LLM extraction per document, with LiteLLM routing and fallbacks
  3. Synthesize — product overview, topic stances, risk score, and rollup across core documents

Deep dives:

License

AGPL-3.0

Contact

About

Transform complex legal documents into clear, actionable insights with AI-powered analysis.

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Sponsor this project

  •  

Contributors