Help & Support Hub

SinAi Support Center

Setup guides, administrator configuration, troubleshooting, and issue reporting for SinAi and the Google Docs Add-on.

Google Docs Add-on: Step-by-Step Setup Guide

Once you have installed SinAi for Google Docs from the Google Workspace Marketplace, follow these steps to start editing:

1

Open Any Google Document

Navigate to docs.google.com and open an existing draft or create a new blank document.

2

Launch from the Extensions Menu

Click Extensions in the top menu bar → select SinAi Document Assistant → click Open Sidebar.

3

Highlight Text & Run AI Tools

Select any Sinhala text paragraph in your document, choose your desired task (Grammar Correction, Headline Generation, 5-Tone Rewrite, or Summary), and click Replace / Insert to update your document instantly.

Workspace Admin Configuration

Google Workspace administrators can deploy SinAi domain-wide across all journalists, editors, and newsroom accounts:

  • Domain Installation: Go to the Google Workspace Admin Console → AppsGoogle Workspace Marketplace appsApps list → click Install app and search for SinAi Document Assistant.
  • Organizational Units (OUs): You can grant access to the entire newsroom domain or target specific organizational units (e.g., Editorial Dept, Sub-Editors).
  • Data Governance: The add-on requires minimal scopes (reading selected document text only) and enforces strict per-user RLS data isolation.

Frequently Asked Questions

Does SinAi support older FM / DL newsroom ASCII fonts?

Yes! SinAi includes an integrated legacy transcoder that decodes ASCII-encoded Sinhala newsroom archives into Unicode before performing linguistic corrections.

Do I need an account to use the Google Docs Add-on?

No, the Add-on works out-of-the-box in anonymous trial mode. If you want cross-device history and elevated rate limits, you can sign in with your SinAi account.

Is my document content used to train public AI models?

No. We adhere to the Google API Services User Data Policy. Your document text is processed in-memory solely to fulfill the requested inference pass and is never used to train generalized models.

Report an Issue / Technical Support

Encountered a bug, unexpected grammatical output, or experiencing technical difficulty? We welcome your feedback: