Federated AI Agent Memory: Moving Distilled Knowledge Between Teacher and Student Machines Across Engineering Teams

By | November 13, 2026
0 Comment

Over the previous installments of this series—from Giving Local AI Assistants Long-Term Memory and Learning While You Sleep: Out-of-Band Memory Dreaming to Hardening Claude Code’s Memory: Mistake Ledgers, Per-Prompt Context, and Verifying the Dreamer Itself and Hardening Local AI Agents: Procedural Skills, Mistake Ledgers, and Real-Time Verification—we solved single-machine amnesia. An AI coding assistant running… Read More: Federated AI Agent Memory: Moving Distilled Knowledge Between Teacher and… »

The Autonomous Homelab: Building a State-of-the-Art Local AI Agent Architecture

By | November 12, 2026
0 Comment

The artificial intelligence industry is currently experiencing a profound architectural divergence. On one side lies the Consumer Chat Wrapper: slick, interactive web UIs powered by multi-trillion-parameter frontier models. Yet despite their raw intelligence, they operate in a state of permanent cognitive amnesia. Every time you open a new tab or restart a session, the model… Read More: The Autonomous Homelab: Building a State-of-the-Art Local AI Agent Architecture »

Supercharging Local AI Agents on Signal: Fast-Paths, Live Typing Indicators, and Session Continuity

By | November 11, 2026
0 Comment

Running a full-scale agentic AI—like Google Antigravity (agy), Claude Code, or a local LLM with tool-use capabilities—directly inside your homelab or developer workstation gives you immense power. You can orchestrate deployments, query infrastructure metrics, check smart home sensors, and run automated diagnostics without exposing any private data or API keys to third-party consumer chat wrappers.… Read More: Supercharging Local AI Agents on Signal: Fast-Paths, Live Typing Indicators,… »

Hardening Claude Code’s Memory: Mistake Ledgers, Per-Prompt Context, and Verifying the Dreamer Itself

By | November 11, 2026
0 Comment

This is the sixth post in my series on giving Claude Code a long-term memory in Obsidian. The earlier posts covered session archiving and context injection, sub-second vault search, self-curating memory, out-of-band “dreaming” and teaching the agent real Obsidian syntax. In Hardening Local AI Agents: Procedural Skills, Mistake Ledgers, and Real-Time Verification I described a… Read More: Hardening Claude Code’s Memory: Mistake Ledgers, Per-Prompt Context, and Verifying… »

Hardening Local AI Agents: Procedural Skills, Mistake Ledgers, and Real-Time Verification

By | November 10, 2026
0 Comment

In Learning While You Sleep: Implementing Anthropic’s Out-of-Band Memory “Dreaming” in Obsidian Part 4 of this series]], we implemented Anthropic’s out-of-band “Dreaming” architecture for my local CLI assistant, Google Antigravity (agy). By running a nightly batch consolidation cycle over raw session transcripts, the agent gained the ability to autonomously distill episodic events into durable declarative… Read More: Hardening Local AI Agents: Procedural Skills, Mistake Ledgers, and Real-Time… »

Teaching Claude Code to Write Real Obsidian Notes: Adding kepano’s obsidian-skills to a Self-Maintained Memory

By | November 9, 2026
0 Comment

This is the fifth post in my series on giving Claude Code a long-term memory in Obsidian. The earlier posts covered session archiving and context injection, sub-second vault search, self-curating memory and out-of-band “dreaming”. Together they let the agent read from my vault and write to it on its own. Something was still missing: the… Read More: Teaching Claude Code to Write Real Obsidian Notes: Adding kepano’s… »

Learning While You Sleep, Part 2: Out-of-Band Memory “Dreaming” for Claude Code

By | November 8, 2026
0 Comment

This is the Claude Code companion to Learning While You Sleep. In that post my agy agent replayed its own sessions overnight and distilled them into its Obsidian memory, without me in the loop. The Claude Code series so far: All three are in-band: they happen during a session, while the agent’s attention is on… Read More: Learning While You Sleep, Part 2: Out-of-Band Memory “Dreaming” for… »

Self-Curating Memory for Claude Code: How the Agent Writes, Deduplicates and Prunes Its Own Obsidian Knowledge Base

By | November 8, 2026
0 Comment

This is the Claude Code companion to Self-Curating AI Memory, where my agy agent started writing to its own knowledge base. Two posts ago I gave Claude Code session archiving and context injection. The previous post added sub-second vault search. Both are about reading: Claude starts each session with my working state, recent sessions and… Read More: Self-Curating Memory for Claude Code: How the Agent Writes, Deduplicates… »

Sub-Second Knowledge Retrieval for Claude Code: Obsidian Omnisearch, a Python Fallback and a Claude Skill

By | November 8, 2026
0 Comment

In the previous post I gave Antigravity (agy) a third memory tier: ranked search over my whole Obsidian vault. It used Obsidian Omnisearch’s hidden HTTP API when Obsidian was running and a standard-library Python scanner when it wasn’t. Since then I have ported the first two tiers to Claude Code: session archiving into SQLite and… Read More: Sub-Second Knowledge Retrieval for Claude Code: Obsidian Omnisearch, a Python… »

Giving Claude Code Long-Term Memory: Session Archiving, Hook-Driven Context Injection, SQLite and Obsidian

By | November 8, 2026
0 Comment

In an earlier post I gave Antigravity (agy) long-term memory. A read-only exporter pulled its SQLite session database into my Obsidian vault, systemd –user units kept that export current, and a five-layer set of notes gave the agent a place to keep state between sessions. I also use Claude Code daily, in the terminal and… Read More: Giving Claude Code Long-Term Memory: Session Archiving, Hook-Driven Context Injection,… »