Skip to content

USAGE.md

Latest commit

 

History

History
158 lines (116 loc) · 3.1 KB

USAGE.md

File metadata and controls

158 lines (116 loc) · 3.1 KB

Usage Guide

CLI Reference

The CLI provides a comprehensive set of commands for managing the memory system.

Status

Check system status:

npx tsx src/cli.ts status
npx tsx src/cli.ts status --json  # JSON output

Recall

Search for relevant memories:

# Basic search
npx tsx src/cli.ts recall "How does authentication work?"

# With options
npx tsx src/cli.ts recall "auth setup" --project api-gateway --limit 5

# JSON output
npx tsx src/cli.ts recall "testing strategy" --json

Options:

  • --project <name> - Filter/boost by project
  • --limit <n> - Maximum memories to return (default: 10)
  • --format <type> - Output format: markdown, xml, plain, compact
  • --json - Return raw JSON

List

List recent memories:

# List last 10 memories
npx tsx src/cli.ts list

# With filters
npx tsx src/cli.ts list --limit 20 --project frontend

Store

Store a conversation for extraction:

npx tsx src/cli.ts store "We decided to use PostgreSQL for the database"
npx tsx src/cli.ts store "..." --project backend

Backup Commands

# Create backup
npx tsx src/cli.ts backup create
npx tsx src/cli.ts backup create --name "before-migration"

# List backups
npx tsx src/cli.ts backup list

# Restore from backup
npx tsx src/cli.ts backup restore ./backups/backup-2024-01-01.tar.gz

Sync Commands

# Check sync status
npx tsx src/cli.ts sync status

# Get setup instructions
npx tsx src/cli.ts sync setup

Config Commands

# Show current config
npx tsx src/cli.ts config show

# Update config
npx tsx src/cli.ts config set retrieval.maxTokens 3000
npx tsx src/cli.ts config set logLevel debug

Interactive REPL

Start the interactive REPL:

npx tsx src/repl.ts

REPL Commands

Command Description
.recall <query> Search for memories
.store <text> Store text for extraction
.list List recent memories
.get <id> Get memory by ID
.delete <id> Delete memory
.status Show system status
.help Show help
.quit Exit

Natural Language Mode

Simply type a question to recall relevant memories:

> How do we handle rate limiting?
[Returns relevant context]

> remember: The API uses Redis for session storage
[Stores for extraction]

Configuration

Configuration is stored in ~/.openclaw-memory/config.json:

{
  "dbPath": "~/.openclaw-memory",
  "enableSync": true,
  "enableAutoExtraction": true,
  "extractionModel": "qwen2.5:7b",
  "embeddingModel": "nomic-embed-text",
  "backup": {
    "autoBackup": true,
    "interval": 24,
    "maxBackups": 10
  },
  "retrieval": {
    "maxTokens": 2000,
    "minScore": 0.5,
    "maxResults": 10,
    "defaultFormat": "markdown"
  },
  "logLevel": "info"
}

Best Practices

  1. Project Organization: Use consistent project names for better context
  2. Importance Scoring: Use 1-3 for low, 4-6 for medium, 7-10 for high
  3. Tags: Add relevant tags for additional filtering
  4. Backups: Create backups before major changes
  5. Sync: Set up Syncthing for cross-device access