Inspiration

We noticed a simple but massive problem: people don’t read what they agree to online.

Existing tools try to summarise Terms & Conditions into plain English, but they remain static, shallow, and disconnected from real-world legal risk patterns.

We wanted to go further, beyond summarisation, and build an AI system that behaves like a real legal reasoning agent, not just a text explainer.

That’s how 5-Second Lawyer was born, an autonomous AI agent that doesn’t just read policies, but actively reasons about risk, compares it across companies, and builds memory over time.


What it does

5-Second Lawyer is a Chrome extension that analyses Terms & Conditions and privacy policies in real time before users accept them.

When activated, an AI agent launches an autonomous reasoning loop:

  1. Fetches and parses the legal document from the active website
  2. Extracts clauses and classifies them into risk categories
  3. Compares them against a persistent graph memory of known legal-risk patterns
  4. Generates a clear, actionable verdict in seconds

Instead of long summaries, users see:

  • “High Privacy Risk Detected”
  • “Auto-Renewal Clause Found”
  • “Third-Party Data Sharing Identified”
  • “Safe to Proceed”

This transforms legal text into instant decision intelligence before consent.


Why this is different

Unlike traditional “AI summarizers,” 5-Second Lawyer is built as a multi-step agentic system.

It doesn’t respond in one shot, it reasons step by step using tools.

The agent dynamically uses:

  • fetch_url
  • risk_analysis
  • check_memory
  • graph_query
  • done

Each step is executed live and visible inside the extension, so users can actually see how the AI reached its conclusion.


How it works

The system is built around three core layers:

1. Agent Runtime (Kimchi)

A model-agnostic execution layer that runs the full agent workflow using OpenAI-compatible or local models, enabling fast and cost-efficient reasoning.

2. Memory Layer (Neo4j)

A persistent graph database that stores:

  • companies
  • clause types
  • privacy risks
  • historical legal patterns

This allows cross-company intelligence, such as:

“This clause matches risk patterns previously observed in major platforms like Uber, Airbnb, and Spotify.”

3. Reasoning Layer (Cypher / Tessl)

A graph-query system that retrieves and updates legal relationships in real time, enabling dynamic risk comparison and visualization during scans.


Impact

Instead of expecting users to understand legal documents, 5-Second Lawyer shifts the responsibility to AI, acting as a real-time trust layer for the internet.

It helps users make faster, safer, and more informed decisions before agreeing to anything online.


Vision

We are building toward a future where every digital agreement is automatically evaluated by intelligent agents before a user commits.

5-Second Lawyer is the first step toward an AI-powered trust infrastructure for the web.

Built With

Share this project:

Updates