Ask Sage Edge
Local AI, shared management workflows, and guidance for every Edge deployment
Table of Contents
About this guide
Ask Sage Edge brings the Ask Sage workspace, document processing, and AI services to infrastructure operated by your organization. With local models and the required services installed, you can chat and work with your data without a public internet connection. Connected deployments can also use approved cloud model providers.
These guides cover everyday use and administration across Ask Sage Edge platforms. The Appliance Manager provides the same core workflows for checking health, managing models, configuring services, and applying updates. Hardware capacity, node count, and physical installation requirements follow your deployment configuration.
Start here
Two places to sign in
Ask Sage workspace is where users chat, upload documents, manage datasets, and use enabled tools. Appliance Manager is where authorized operators manage the appliance and its models. A workspace account does not automatically grant appliance administration rights.
The standard Edge workspace address is https://chat.edge.asksage.ai. Your organization may assign its own domain. These names must resolve to your deployment on your network; they are not a shared public Edge service. Obtain your Manager address and credentials from your administrator.
What works without the internet?
Local chat, local embedding and retrieval, and installed local speech services can work offline. Cloud models, external search, remote connectors, and internet-hosted identity providers require access to their respective services. A model’s available tools and supported input types also determine which workspace features work.
Selecting Local inference changes the model configuration. Network isolation and restrictions on external integrations remain part of your organization’s deployment configuration. See offline operation before disconnecting.
Match the guide to your installation
Use these Manager instructions across Edge platforms. Screenshots show an example appliance; its hardware label, node count, models, and status readings do not define a required configuration. Controls depend on your role, installed release, and enabled capabilities. Follow your deployment handover for hardware installation and recovery details.