Guardrails (Staging)

About

Guardrails is a privacy-first AI assistant for nonprofits and mission-driven teams.
It helps you use tools like ChatGPT, Claude, and Google Gemini safely, without exposing sensitive information.

Why Guardrails?

Teams are increasingly using AI to draft emails, analyse reports, and summarise casework. But copying real data into AI tools can expose personal, donor, or beneficiary information.

Guardrails solves this by acting as a protective layer between you and AI tools, directly in your browser.

What it does

1. Detects sensitive data instantly
Scans what you type or paste into AI tools and identifies personal or high-risk information.

2. Automatically redacts or anonymises
Masks names, emails, case details, and other sensitive data before it is sent.

3. Explains what and why
Shows exactly what was changed and gives clear, simple explanations so you understand the risks.

4. Teaches safe AI use as you work
Provides short tips and micro-lessons on topics like GDPR, privacy, and responsible AI.

5. Works with your existing tools
Use your preferred AI platforms and API keys while applying consistent protection rules.

Key features
Real-time PII detection and redaction
Multiple modes: mask, pseudonymise, or block
Transparent “before and after” view of prompts
Lightweight and fast, no complex setup
Built-in learning prompts and guidance
Designed for nonprofits, NGOs, and ethical teams

Who it’s for
Nonprofits and charities
Social enterprises and cooperatives
Teams handling sensitive or personal data
Anyone who wants to use AI more responsibly

Your data stays protected

Guardrails is built around privacy and dignity:

Data is processed locally or securely, depending on setup
No data is used to train external AI models
Clear visibility into what happens to your data

Our mission

We believe organisations should be able to use AI without compromising the people they serve. Guardrails helps teams adopt AI with confidence, while protecting privacy and building real understanding.