Vanar Neutron Memory

Vanar Neutron Memory

Persistent memory storage with semantic search for AI agents. Save text as seeds, search semantically, and persist agent context between sessions.

Features

• Auto-Recall: Automatically queries relevant memories before each AI turn and injects as context
• Auto-Capture: Automatically saves conversations after each AI turn
• Semantic Search: Find memories by meaning using Jina Embeddings v4 (1024 dimensions)
• Memory Types: Episodic, semantic, procedural, and working memory
• Blockchain Attestation: Tamper-evident memory storage with transaction hashes

Prerequisites

Get your API keys at: https://openclaw.vanarchain.com/

API credentials via environment variables:

export NEUTRON_API_KEY=your_key
export NEUTRON_AGENT_ID=your_agent_id
export YOUR_AGENT_IDENTIFIER=your_agent_name_or_id  # agent_id name or defaults to 1

Or stored in ~/.config/neutron/credentials.json:

{
  "api_key": "your_key_here",
  "agent_id": "your_agent_id_here",
  "your_agent_identifier": "your_agent_name_or_id"
}

Testing

Verify your setup:

./scripts/neutron-memory.sh test  # Test API connection

Hooks (Auto-Capture & Auto-Recall)

The skill includes OpenClaw hooks for automatic memory management:

• hooks/pre-tool-use.sh - Auto-Recall: Queries memories before AI turn, injects relevant context
• hooks/post-tool-use.sh - Auto-Capture: Saves conversation after AI turn

Configuration

Both features are enabled by default. To disable:

export VANAR_AUTO_RECALL=false   # Disable auto-recall
export VANAR_AUTO_CAPTURE=false  # Disable auto-capture

Or add to your credentials file:

{
  "api_key": "your_key_here",
  "agent_id": "your_agent_id_here",
  "your_agent_identifier": "your_agent_name_or_id",
  "auto_recall": true,
  "auto_capture": true
}

Scripts

Use the provided bash script in the scripts/ directory:

• neutron-memory.sh - Main CLI tool

Common Operations

Save Text as a Seed

./scripts/neutron-memory.sh save "Content to remember" "Title of this memory"

Semantic Search

./scripts/neutron-memory.sh search "what do I know about blockchain" 10 0.5

Create Agent Context

./scripts/neutron-memory.sh context-create "my-agent" "episodic" '{"key":"value"}'

List Agent Contexts

./scripts/neutron-memory.sh context-list "my-agent"

Get Specific Context

./scripts/neutron-memory.sh context-get abc-123

Interaction Seeds (Dual Storage)

When NeutronMemoryBot processes an interaction, it stores data in two places:

1. Agent Context - Truncated summary for structured metadata and session tracking
2. Seed - Full thread snapshot for semantic search

Each time the bot replies to a comment, the full thread (original post + all comments + the bot's reply) is saved as a seed. This means:

• Every seed is a complete conversation snapshot
• Later seeds contain more context than earlier ones
• Semantic search finds the most relevant conversation state
• Append-only: new snapshots are added, old ones remain

Seed Format

Thread snapshot - {timestamp}

Post: {full post content}

Comments:
{author1}: {comment text}
{author2}: {comment text}
NeutronMemoryBot: {reply text}

API Endpoints

• POST /seeds - Save text content (multipart/form-data)
• POST /seeds/query - Semantic search (JSON body)
• POST /agent-contexts - Create agent context
• GET /agent-contexts - List contexts (optional agentId filter)
• GET /agent-contexts/{id} - Get specific context

Auth: All requests require Authorization: Bearer $NEUTRON_API_KEY header and appId/externalUserId query params.

Memory types: episodic, semantic, procedural, working

Text types for seeds: text, markdown, json, csv, claude_chat, gpt_chat, email