Agent Brain 🧠

Teach your AI once. It remembers forever. It gets smarter over time.

Agent Brain is a modular memory system for AI agents with continuous learning. It stores facts, catches contradictions, learns your habits, ingests external knowledge, tracks what works, learns from mistakes, and adapts to your tone — all in a local SQLite database with real persistence, full-text search, and pluggable storage backends.

Why this exists

Every AI conversation starts from zero. You repeat yourself. It forgets what you taught it. Agent Brain fixes that with a working persistence layer (scripts/memory.sh) and six cognitive modules that the agent selectively invokes based on what the task actually needs.

What makes this different

• Production-grade storage. SQLite with WAL mode and indexed queries. Handles 10,000+ entries without breaking a sweat. JSON backend available as fallback.
• Pluggable backends. Storage abstraction layer means you can swap SQLite for Postgres, Supabase, or any other backend — the command interface stays the same.
• Continuous learning. Corrections track what was wrong, what's right, and why. Successes reinforce what works. Anti-patterns emerge from repeated mistakes.
• Selective dispatch, not a linear pipeline. The orchestrator picks the 1-3 modules relevant to each task. Storing a fact doesn't need Vibe. Reading tone doesn't need Archive.
• Active fact extraction. The agent scans every message for storable information — identity, tech stack, preferences, workflows, project context — without being asked.
• Honest confidence. No fake 0.0-1.0 scores. Four categories (SURE / LIKELY / UNCERTAIN / UNKNOWN) derived from actual metadata — source type, access count, age.
• Hybrid retrieval. Results ranked with lexical + semantic scoring (--policy fast|balanced|deep) and optional score explainability (--explain).
• Supersede, don't delete. Old facts aren't destroyed. They're marked superseded_by with a pointer to the replacement, preserving full history.
• Decay is mechanical. Entries scale their decay threshold by access count. Heavily-used knowledge persists longer. Unused knowledge fades.

Architecture

Six modules, one orchestrator, pluggable storage.