Aegistra — AI governance for small teams

AI shows up faster than anyone writes it down.

Someone adds a chatbot to support. Sales trials a writing tool. A vendor ships an AI feature in an app you already pay for. Then a customer's security questionnaire asks which AI systems you use, who owns them and when they were last reviewed.

Aegistra keeps those answers in one register, so the next questionnaire is a lookup instead of a scramble.

Which AI systems do you use?AI systems
Who is responsible for each one?Owner
When was it last reviewed?Review due

What's in Aegistra today

Every screen below is in the live app. We filled them with sample data.

One register for every AI system

Add each AI tool, model, agent or AI feature your team relies on. Every entry records what it is for, who owns it and how it is used.

System name, provider and purpose
Owner name and email
Lifecycle: pilot, production, paused or retired
Next review date
Data sensitivity, autonomy and potential impact

/app/systemsSample data

AI systems

Every AI-enabled system or business use your team relies on.

System	Owner	Lifecycle	Priority	Review due
Candidate screening copilot Recruiting vendor · Ranks inbound job applicants	Priya Nair	Production	high · 75	2026-10-01
Contract clause finder Legal research tool · Finds clauses in vendor contracts	Ana Ruiz	Pilot	medium · 35	2026-12-10
Meeting summarizer Built into a meeting app · Summarizes internal calls	—	Production	low · 18	Not set
Support reply agent Model API · Drafts replies to customer tickets	Marcus Lee	Production	medium · 49	2026-12-01

See what needs attention first

Your overview counts your systems, your high-priority systems, the reviews that are due and the actions still open. The riskiest systems come first.

Systems sorted by priority score
Reviews that are due, counted for you
Open actions with an owner and a due date

/appSample data

Overview

Northwind Studio

A living view of your AI systems, reviews and governance work.

AI systems

High priority

Reviews due

Open actions

Systems needing attention

Candidate screening copilot
Recruiting vendor · Priya Nair
high · 75
Support reply agent
Model API · Marcus Lee
medium · 49
Contract clause finder
Legal research tool · Ana Ruiz
medium · 35

Open actions

Review vendor data retention terms
Priya Nair · due 2026-10-15
Add human sign-off to candidate shortlists
Priya Nair · due 2026-10-20
Assign an owner to Meeting summarizer
Ops · due 2026-10-25

Turn findings into tasks

Write down what needs to happen, such as reviewing a vendor's data retention terms. Give it an owner and a due date, and mark it done when it is finished.

An owner and a due date on every action
Mark an action done in one click
Everything open in one list

/app/actionsSample data

Workflow

Governance actions

Add action

Review vendor data retention terms

Owner

Due date

Create action

Review vendor data retention terms
Priya Nair · due 2026-10-15
Mark done
Add human sign-off to candidate shortlists
Priya Nair · due 2026-10-20
Mark done
Assign an owner to Meeting summarizer
Ops · due 2026-10-25
Done

Find the gaps before someone asks

The readiness view shows which systems have an owner and a review date and which do not, so you can fix gaps before a customer or auditor finds them.

Owner: ready or missing
Review date: set or missing
Priority level next to every system

Download the whole register, reviews, actions and evidence as a PDF assurance pack or CSV files from the Evidence page.

/app/evidenceSample data

Evidence

Assurance readiness

Spot missing ownership and review evidence.

Registered systems

With owner

With review date

System	Owner	Review date	Priority
Candidate screening copilot	Ready	2026-10-01	high
Contract clause finder	Ready	2026-12-10	medium
Meeting summarizer	Missing	Missing	low
Support reply agent	Ready	2026-12-01	medium

Also in the app

System detail and edit: Open any system to change its answers, owner or review date. The score updates with it.
Review history: Record each review as approved, changes required or paused, with notes and the next due date.
Team invitations and roles: Invite teammates as admin, member or viewer. Viewers can read but not change anything.
Private evidence: Attach links, notes, decisions or files (PDF, images, Office documents, up to 15 MB) to a system. Files are stored privately.
CSV import and export: Download a template, import many systems at once, or export the whole register.
Activity log: See who added, changed or removed systems, actions, evidence and members, and when.

A score you can explain

Each system is scored from 0 to 100 with six questions you answer in plain language.

The weights are fixed and published here, and the demo at the top of this page runs on them. You can always see exactly why a system ranks where it does.

The score ranks your internal review work. It is not a legal risk classification, and it does not certify compliance with any regulation.

Data sensitivity

none: 0
internal: +8
personal: +18
sensitive: +28

Autonomy

assistive: +3
recommendation: +10
decision: +22
autonomous: +30

Potential impact

low: +4
moderate: +14
high: +25

Added when it applies

No human review before action: +10
Talks to customers or the public: +4
Generates visible content: +3

Priority levels

Low: 0 to 34
Medium: 35 to 64
High: 65 to 100

From a blank register to a current one

Step 1
Create a workspace and add your AI systems
Name your workspace, then add each system with its provider, purpose, owner and lifecycle.
Step 2
Answer six questions for each system
Data sensitivity, autonomy, potential impact, human review, customer contact and visible content. Aegistra turns your answers into a priority score.
Step 3
Assign reviews and close the gaps
Set a next review date, add actions, and use the overview and readiness view to see what is overdue or missing.

Built for teams without a GRC department

Aegistra is for software companies, agencies and consultancies with roughly 10 to 250 people, where customers increasingly ask how you use AI.

Who usually owns this

Founders and COOs
Security leads
Privacy and compliance leads
IT leads
Fractional CISOs and consultants

Start free. Upgrade when you outgrow it.

Three AI systems are free. Team is $5 a month and Business is $10 a month. Cancel any time and keep access until the period ends.

Plan comparison
Feature	Free	Team	Business
Price	$0	$5per month	$10per month
AI systems	3	50	Unlimited
Assurance pack (PDF and CSV)	Included	Included	Included
Availability	Available now	Available now	Available now

Every plan includes the register, priority scoring, overview, governance actions and readiness view. The plans differ in how many AI systems you can add.

Create a free workspace Upgrade an existing workspace

Prices are in US dollars. Payments are handled by Razorpay. Only the workspace owner can change the plan.

We store governance metadata, not your prompts.

Aegistra only needs to know what a system is and who looks after it.

No model traffic: Aegistra is not in the path between your apps and AI providers. It records what a system is, who owns it and when it was reviewed. It never sees your prompts, outputs or the data you send to models.
Isolated workspaces: Every record belongs to one workspace. Access is enforced by row-level security in the Postgres database, not only by what the screens show.
Keys stay on the server: Your browser only receives a publishable key. The secret key used by the admin area never leaves the server.
Hosted on trusted infrastructure: Aegistra is served over HTTPS from Vercel. Sign-in and data are handled by Supabase.

Where the product is headed

Aegistra is early. Here is what works today and what we are building next.

Live now

Sign-up and workspace setup
AI system register
Priority scoring
Overview dashboard
Governance actions
Readiness view
Isolated workspaces
System detail and edit screens
Review history
Team invitations and roles
Private evidence uploads
CSV import and export
Activity log
Assurance pack export (PDF and CSV)
Paid plans with Razorpay billing

Review reminder emails (built, switching on soon)

After that

Answer library for security questionnaires
Framework references and control mapping
Security hardening for paid launch

Later

Google Workspace and Microsoft 365 discovery experiments
Slack and Teams workflow
Employee requests to approve AI tools
Multi-client mode for consultants and MSPs

The order can change as we learn what early teams need. Planned items are not commitments.

What Aegistra does not do

It is not legal advice, a certification or a legal risk classification.
It does not watch your AI traffic or prompts.
It does not discover AI tools for you. You add the systems you know about.
It does not offer single sign-on or framework mapping yet.

Questions teams ask first

What counts as an AI system?

Anything that uses AI to do work for your business: a chatbot, a model API call inside your product, an agent, a writing assistant, or an AI feature inside software you already pay for.

Does Aegistra see our prompts or customer data?

No. It does not sit between your apps and AI providers. You enter a description of each system and answer six questions about it.

Is the score a legal risk classification?

No. It ranks your own review work using the six answers you give. Deciding how a system is classified under any AI regulation stays with you and your advisers.

Can we invite teammates or import a spreadsheet?

Yes. Owners and admins can invite teammates as admin, member or viewer, and you can import systems from a CSV file or export your register as CSV.

What does it cost?

The Free plan covers your first 3 AI systems. Team is $5 a month for up to 50 AI systems and Business is $10 a month with no limit. Pay by card or other methods Razorpay supports, and cancel any time.

Add your first AI system today.

Create a free workspace, add up to three systems and see their priority scores.

Create a free workspace Log in

No credit card.

Know where AI is used.Know who owns it.Be ready when customers ask.

Add an AI system

Where the points come from