
Summaries like this, in your inbox every morning.
Sign up free →Patina ingests Slack exports and builds a persistent belief graph—tracking entities, relationships, and claims with confidence decay—to learn your priorities, communication style, and decision patterns without requiring universal rule configuration.
The system operates across three tiers (deterministic scoring, local LLM via Qwen 3.x or Ollama, and frontier LLM via Claude or GPT-4o), staying local-first with all data stored in ~/.patina/store.db; users can run Tier 1 alone with zero LLM calls and graduate autonomy levels (0–6) earned by accuracy rather than configured.
Day 1 delivers a unified catch-up view; by Day 30 the tool drafts in your voice and dismisses noise; by Day 90 it predicts actions, flags contradictions across conversations, and operates silently—available now via `uv tool install patina` with Apache 2.0 licensing.
No discussion yet for this article
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
5 minutes a day. The AI essentials.
200+ sources · Email / LINE / Slack