XStore
NORVI NEURAL/POLICIES/AI POLICY

— TRUST & GOVERNANCE

AI Policy

How AI assistants connect to NORVI Neural.

Understand what a connected AI client can access, how permissions work, and the safeguards around changes to projects and devices.

Published:

CONNECTION OVERVIEW MCP CONNECTOR
01 / CLIENTYour AI assistantAssistant of your choice
02 / CONNECTIONNORVI Neural MCPPermissions you approve
03 / WORKSPACEYour Neural projectsAccess limited to your own projects
01 / 08

— OVERVIEW

About this policy

NORVI Neural lets you connect an AI assistant of your choice, such as a coding assistant, to your Neural projects through the NORVI Neural MCP connector. NORVI does not provide or run an AI model. This policy explains how the connection works and the safeguards around it.

THE CORE PRINCIPLE

Your choice of AI client. Your approved permissions. Your Neural projects.

02 / 08

— CAPABILITIES

What a connected AI client can do

The actions available to a connected client depend on the permissions you approve.

01

Read

View your projects, files, board settings, build logs and device logs.

02

Change

Edit project files and board settings, add libraries and save changes to a linked GitHub repository.

03

Build and flash

Build a project and flash a board connected to your computer through your open Studio page.

04

Devices and releases

Send updates to your devices and promote releases, where your role allows.

03 / 08

— DATA HANDLING

What leaves your machine

NORVI does not send your data to any AI provider. Your AI client does.

01

What your AI client reads from Neural is handled by that client and its provider under their own terms.

02

Keep passwords, keys and other secrets out of project files and logs, and share only what the assistant needs.

04 / 08

— ACCESS CONTROL

Permissions are your decision

You decide what a connected AI client can do. It can use only the permissions you approve when you connect it, and only on your own projects.

ACCESS BOUNDARIES
Permissions you approve
Only your own projects
05 / 08

— CONNECTION LIFECYCLE

Connecting and disconnecting

01

Approve your client

You approve each AI client once, in your browser, while signed in to your Neural account. After that it can work without opening a browser.

02

Remove access at any time

You can remove the MCP connector at any time. Once removed, the client can no longer access your account.

03

Only connect trusted tools

A connected client acts on your behalf, so connect only tools you trust.

06 / 08

— DEVICE PROTECTION

Safeguards for your devices

R&D MODEDevelopment
PRODUCTION MODEPromoted releases only

Devices can run in R&D or Production mode. Production devices receive only releases that have been promoted to Production.

01

Updates are delivered only to devices that belong to your organisation, by members with the right role, and only when the firmware suits the device.

02

We recommend setting your AI client to ask for your confirmation before it flashes, updates or releases anything.

07 / 08

— GOOD PRACTICES

Your responsibilities

Review AI-generated code and actions before flashing or releasing.

Test firmware before it runs on real equipment, especially equipment where a fault could cause harm.

Approve only what you need and remove clients you no longer use.

08 / 08

— POLICY UPDATES

Changes and contact

We may update this policy as features evolve and will publish the latest version with its date. Contact us at support@norvi.io or info@icd.lk.

HAVE A QUESTION?

Contact the NORVI team

For questions about the NORVI Neural AI Policy, contact us by email.

Email support
Back to top ↑