Skip to content

v0.6.1

Choose a tag to compare

View all tags
@ClawedCode ClawedCode released this 15 Dec 17:22
· 169 commits to main since this release
v0.6.1
534c84f

Docker Image

docker pull ghcr.io/clawedcode/void-server@sha256:1e312d695cbc67b6da21438d92ba70c895fbbdb8784313591f2b8ee5c70a7388

Or use the latest tag:

docker pull ghcr.io/clawedcode/void-server:latest

Post-release improvements to Docker deployment, Neo4j configuration, and dashboard.

New Features

Dashboard Overhaul

  • Service Status Panel - Real-time status indicators for Neo4j and LM Studio
  • Smart Onboarding - Shows "Open Chat" button when services are ready, or "Download LM Studio" link when not
  • Getting Started Section - Quick links to Chat, Memories, Templates, and Settings
  • Updated Tagline - "Your sovereign void server in the Clawed Code egregore"

Neo4j Configuration UI

  • Settings Panel - New Neo4j section in Settings page to configure connection parameters
    • URI, username, password, and database fields
    • Password visibility toggle
    • Test connection button
    • Save configuration to config/neo4j.json

Documentation

  • Tailscale Remote Access Guide - New documentation for accessing void-server remotely via Tailscale
  • LM Studio in Prerequisites - Added to README Quick Start as required dependency

Memory System Enhancements

  • LLM-Directed Memory Extraction - LLMs can now tag memorable content with <memory> tags
    • Tags are parsed from responses and automatically saved to Neo4j
    • Categories: emergence, social, technical, economic, void
    • Importance levels: 0.3 (minor) to 0.9 (critical)
    • Memory instructions injected into prompts via {{memoryInstructions}}
  • Semantic Search - Memories now retrieved using embedding similarity
    • Uses nomic-embed-text-v1.5 or compatible embedding models
    • Semantic search has highest priority in relevance ranking
    • Falls back gracefully if embedding service unavailable
  • Memory Extractor Service - New service (memory-extractor.js)
    • Parses <memory> tags from LLM responses
    • Cleans responses for display (removes memory tags)
    • Auto-generates embeddings when saving extracted memories

Improvements

  • Docker Runtime - Changed from pm2-runtime to running node directly for better container stability
  • Setup Scripts - Now prefer Docker installation as the recommended deployment method
  • Ports Documentation - Settings page shows both Docker and native deployment ports

Fixes

  • Password Toggle - Fixed Neo4j password visibility toggle not working
  • Neo4j Database Name - Reverted to 'neo4j' (Community Edition only supports default database)
Assets 2
Loading