Private Chat
Last updated
Was this helpful?
Private Chat is Flashgate’s text-to-text conversational interface for secure interaction with AI models connected through the Cloud and AI Gateway.
It is designed for organizations that want centralized governance, repository-scoped AI access, and usage observability.
Private Chat is a standard multi-turn chat experience where users can:
select an environment,
prompt available AI models,
continue context-aware conversations,
and monitor usage over time.
Private Chat depends on your gateway setup. Before using it, you must have:
an AI/LLM resource configured,
attached to a repository,
and that repository inside a workspace.
Without this chain (workspace → repository → AI resource), Private Chat cannot execute model calls.
Before starting a conversation, users select an environment that maps to:
Workspace + Repository
This determines:
which providers/models are available,
which policies are enforced,
and which permissions apply.
Private Chat exposes models from the providers connected to the selected repository.
Example:
if an OpenAI resource is connected to the selected repository, users can access OpenAI models available through that configured endpoint.
This same logic applies to other connected providers.
Private Chat consumes underlying provider resources through your Flashgate setup. Depending on your configuration and credits, organizations can optimize cost by routing usage through preferred providers or existing cloud credits.
Private Chat includes quick usage visibility (for example, recent consumption windows such as last 7 days), while organization-wide analytics are available in the Organization → Statistics pages.
Private Chat usage is governed by policy scopes defined in the platform:
organization,
workspace,
repository.
This allows privacy and compliance controls to apply consistently to conversational traffic.
Private Chat currently supports core conversation workflows and model interaction. Roadmap-oriented capabilities include:
longer-term memory behavior,
deeper storage-backed memory,
richer agent-based interactions across conversations.
See: Conversation API
Last updated
Was this helpful?
Was this helpful?