BrainX V6

What problem does it solve?

OpenClaw agents lack long-term, searchable memory across sessions, so repeated decisions, gotchas, and operational learnings get lost and cannot be reliably recalled later.

Core Features & Use Cases

• Persistent semantic memory: Stores and retrieves curated memories using PostgreSQL + pgvector embeddings.
• Retrieval + context injection: Performs semantic search and formats selected memories for prompt injection.
• Governance and safety-first recall: Uses verification states, sensitivity controls, and conservative rollout of runtime surfaces.
• Maintenance automation: Supports lifecycle management, deduplication, contradiction handling, diagnostics (doctor), and backup/restore.

Quick Start

Run this to store a new durable memory for a deployment context: brainx add "Use pgvector cosine similarity for semantic recall" --type decision --context project:example --importance 8.