Repository navigation
LibreChat - OIDC Identity, Browser History and Private AI Integration #18048
Petertfcoo
started this conversation in
General
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hello,
We're assessing LibreChat as a lightweight conversational interface for a repeatable, client-owned AWS AI architecture.
The underlying system already provides Amazon Bedrock, private knowledge retrieval, authenticated access controls and authoritative operational audit records. We are particularly interested in LibreChat's database-free approach.
Could you clarify six points?
Identity propagation: Can LibreChat forward an authenticated user's original, verifiable OIDC access token to a configured OpenAI-compatible endpoint without source-code changes?
Conversation history: Can an administrator enforce a maximum retention period for conversations stored in browser IndexedDB, including clearing history on logout or session expiry?
Knowledge changes: Does LibreChat support clearing or invalidating conversational context when an external middleware service reports that a policy or SOP has been superseded?
Feature restrictions: Can administrators disable optional MCP tools, alternative model endpoints, external connections and user-defined routing through supported configuration?
Audit and response metadata: Can LibreChat pass authenticated user/session identifiers and display structured response metadata, including citations, rule status, escalation indicators and controlled document links, without custom development?
Deployment and support: Is LibreChat intended for production deployment across independent organisations, and what arrangements exist for security updates, maintenance and long-term support?
Our objective is a secure, straightforward conversational interface that an SME IT function or appointed AWS partner can deploy and maintain without a custom application fork.
Please distinguish existing documented functionality from capabilities that would require modification.
Many thanks.
All reactions