Skip to content

Gemini native context caching for the foundational prompt #9

Description

@MuddyWinds

Use Gemini's native context caching to cache the foundational/system prompt across requests, cutting token cost and latency at scale. The enterprise fork targets 200+ channels, where re-sending the foundational prompt on every request is expensive.

Constraint: cache TTL and cache-key management must align with the batching pipeline. Enterprise-only; pairs with the dynamic Situational Context prompt as the non-cached portion.

Metadata

Metadata

Assignees

No one assigned

    Labels

    enhancementNew feature or requestenterpriseEnterprise fork feature

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions