A post-retrieval temporal layer for RAG systems — validity filtering, time decay, document kind classification, and hybrid reranking in one pipeline.
-
Updated
May 15, 2026 - Python
A post-retrieval temporal layer for RAG systems — validity filtering, time decay, document kind classification, and hybrid reranking in one pipeline.
Validate retrieved context before AI agents act.
This repository contains the coding materials to reproduce the learning error curves of the paper "How Does Data Freshness Affect Real-time Supervised Learning ?".
Markdown vault management toolkit -- freshness, refactoring, indexing, and tags. Zero API dependency CLI for Obsidian vaults and knowledge bases.
Keeps documentation honest with the code it describes — four axes, one creed, prove it or drop it: a ride-along reflex flags drift, winnow certifies every claim in a deep pass, flourish crafts ugly-but-true docs to a gold standard, cultivate keeps the repo clean, and seed writes code-backed docs where none exist, only with the owner's approval.
Temporal-awareness hook for tool-using LLMs — surfaces freshness verdicts on in-context data so stale information doesn't poison fresh reasoning. Adapters for Claude Code, Antigravity, and any generic harness.
Make a derived artifact refuse to pretend it is fresh.
MLOps feature quality, freshness, offline/online consistency, lineage evidence, CLI/API gates, Prometheus, Kubernetes, and Terraform
Tell whether your data is fresh, not merely rewritten. Four independent freshness layers -- the signal is the divergence between them.
Your scheduled job says exit 0 — prove it did the work. Three stdlib-only checks: output freshness, silent no-op detection, rollout proof by reading the fact back. Sanitized from a live agent fleet. Free, MIT.
Add a description, image, and links to the freshness topic page so that developers can more easily learn about it.
To associate your repository with the freshness topic, visit your repo's landing page and select "manage topics."