Skip to content

CLI Cheatsheet

Jump to bottom
Varun Pratap Bhardwaj edited this page Mar 16, 2026 · 1 revision

CLI Cheatsheet

Quick reference for all SuperLocalMemory commands - Copy-paste ready commands organized by task for fast lookup.

Memory Operations

Remember (Save)

# Basic memory
slm remember "content"

# With tags
slm remember "content" --tags tag1,tag2,tag3

# With project
slm remember "content" --project myapp

# With importance (1-10)
slm remember "content" --importance 8

# All options combined
slm remember "Critical: Production requires approval" \
  --tags deployment,production \
  --project myapp \
  --importance 10

Parameters:

  • content - Text to remember (required)
  • --tags - Comma-separated tags (optional)
  • --project - Project name (optional, default: "default")
  • --importance - Priority 1-10 (optional, default: 5)

Recall (Search)

# Basic search
slm recall "query"

# Limit results
slm recall "query" --limit 5

# Minimum relevance score (0.0-1.0)
slm recall "query" --min-score 0.7

# Filter by tags
slm recall "query" --tags security,auth

# Filter by project
slm recall "query" --project myapp

# Combined filters
slm recall "authentication" \
  --limit 3 \
  --min-score 0.7 \
  --tags security \
  --project myapp

Parameters:

  • query - Search query (required)
  • --limit - Max results (optional, default: 10)
  • --min-score - Minimum relevance 0.0-1.0 (optional, default: 0.3)
  • --tags - Filter by tags (optional)
  • --project - Filter by project (optional)

List Recent

# List 10 most recent
slm list

# Custom limit
slm list --limit 20

# Filter by project
slm list --project myapp

# Filter by tags
slm list --tags important

Parameters:

  • --limit - Number to show (optional, default: 10)
  • --project - Filter by project (optional)
  • --tags - Filter by tags (optional)

Knowledge Graph

Build Graph

# Build or update graph
slm build-graph

# Force complete rebuild
slm build-graph --force

# With topic clustering (requires python-igraph, leidenalg)
slm build-graph --clustering

# Verbose output
slm build-graph --verbose

# Dry run (preview without saving)
slm build-graph --dry-run

When to run:

  • After bulk imports (50+ memories)
  • Monthly maintenance
  • When search quality degrades
  • After database restore

See Knowledge Graph Guide for details.

Graph Statistics

# View graph stats
slm status

# Detailed graph info
slm status --verbose

Profile Management

Switch Profiles

# List all profiles
slm switch-profile list

# Create new profile
slm switch-profile create work --description "Work projects"

# Switch to profile
slm switch-profile work

# Switch back to default
slm switch-profile default

# Delete profile (with confirmation)
slm switch-profile delete work --confirm

Use cases:

  • Separate work/personal memories
  • Client-specific knowledge bases
  • Project-specific contexts
  • Experimentation sandboxes

See Multi-Profile Workflows for best practices.

System Management

Status Check

# Basic status
slm status

# Verbose with all details
slm status --verbose

# JSON output (for scripts)
slm status --format json

Status includes:

  • Total memories count
  • Database size
  • Knowledge graph statistics
  • Pattern learning stats
  • Current active profile

Pattern Learning

# Update learned patterns
python3 ~/.claude-memory/pattern_learner.py update

# Get identity context
python3 ~/.claude-memory/pattern_learner.py context 0.5

# View all patterns
python3 ~/.claude-memory/pattern_learner.py list

# Reset patterns
python3 ~/.claude-memory/pattern_learner.py reset

See Pattern Learning Explained for details.

Advanced Operations

Database Management

# Check database integrity
sqlite3 ~/.claude-memory/memory.db "PRAGMA integrity_check;"

# Count total memories
sqlite3 ~/.claude-memory/memory.db "SELECT COUNT(*) FROM memories;"

# List all tags
sqlite3 ~/.claude-memory/memory.db "SELECT DISTINCT tag FROM memory_tags ORDER BY tag;"

# Vacuum database (optimize space)
sqlite3 ~/.claude-memory/memory.db "VACUUM;"

# Backup database
cp ~/.claude-memory/memory.db ~/.claude-memory/memory.db.backup

Export/Import

# Export all memories to JSON
sqlite3 ~/.claude-memory/memory.db \
  "SELECT json_group_array(json_object(
    'id', id,
    'content', content,
    'tags', tags,
    'project_name', project_name,
    'created_at', created_at
  )) FROM memories;" > memories.json

# Import from JSON (Python script)
python3 << 'EOF'
import json, sqlite3
conn = sqlite3.connect("/Users/$(whoami)/.claude-memory/memory.db")
with open('memories.json') as f:
    data = json.load(f)
    for mem in data:
        conn.execute(
            "INSERT INTO memories (content, tags, project_name) VALUES (?, ?, ?)",
            (mem['content'], mem['tags'], mem['project_name'])
        )
conn.commit()
EOF

Bulk Operations

# Bulk import from text file (one per line)
while IFS= read -r line; do
  slm remember "$line"
done < memories.txt

# Bulk import from CSV
while IFS=',' read -r content tags project importance; do
  slm remember "$content" \
    --tags "$tags" \
    --project "$project" \
    --importance "$importance"
done < memories.csv

# Bulk tag update (SQL)
sqlite3 ~/.claude-memory/memory.db \
  "UPDATE memories SET tags = tags || ',reviewed' WHERE created_at < date('now', '-30 days');"

Scripting & Automation

Daily Standup Helper

#!/bin/bash
# Save as: ~/bin/standup.sh

echo "Yesterday's Decisions:"
slm recall "decided" --limit 5

echo -e "\nCurrent Blockers:"
slm recall "blocked" --tags critical --limit 3

echo -e "\nRecent TODOs:"
slm recall "TODO" --limit 5

Git Post-Commit Hook

#!/bin/bash
# Save as: .git/hooks/post-commit

commit_msg=$(git log -1 --pretty=%B)
commit_hash=$(git log -1 --pretty=%H)

slm remember "Commit: $commit_msg (${commit_hash:0:7})" \
  --tags git,commit \
  --project "$(basename $(git rev-parse --show-toplevel))"

Weekly Graph Rebuild (Cron)

# Add to crontab (crontab -e)
# Every Sunday at 3 AM
0 3 * * 0 /usr/local/bin/slm build-graph --clustering >> /var/log/slm-build.log 2>&1

Context Injection for Aider

# Use aider-smart wrapper (auto-context injection)
aider-smart

# Or manually inject context
context=$(slm recall "current project" --limit 3 --min-score 0.7)
aider --message "Context: $context. Now help me with..."

Output Formatting

JSON Output

# Status as JSON
slm status --format json

# Search results as JSON
slm recall "query" --format json

# Parse with jq
slm recall "FastAPI" --format json | jq '.results[0].content'

CSV Output

# Export search results to CSV
slm recall "query" --format csv > results.csv

# Import into spreadsheet
# File → Import → CSV

Common Tasks

Setup New Project

# Create project profile
slm switch-profile create myproject

# Add initial context
slm remember "Tech stack: FastAPI, PostgreSQL, React" --project myproject
slm remember "Repository: github.com/me/myproject" --project myproject
slm remember "Team: @alice, @bob, @charlie" --project myproject

# Build initial graph
slm build-graph

Search by Time Range

# Last 7 days (SQL)
sqlite3 ~/.claude-memory/memory.db \
  "SELECT id, content, created_at FROM memories WHERE created_at >= date('now', '-7 days') ORDER BY created_at DESC;"

# Specific date range
sqlite3 ~/.claude-memory/memory.db \
  "SELECT id, content FROM memories WHERE created_at BETWEEN '2026-02-01' AND '2026-02-07';"

Find High-Importance Memories

# Importance >= 8
sqlite3 ~/.claude-memory/memory.db \
  "SELECT id, content, importance FROM memories WHERE importance >= 8 ORDER BY importance DESC;"

Archive Old Memories

# Export old memories (90+ days)
sqlite3 ~/.claude-memory/memory.db \
  "SELECT * FROM memories WHERE created_at < date('now', '-90 days');" > old_memories.sql

# Delete old memories (with backup first!)
cp ~/.claude-memory/memory.db ~/.claude-memory/memory.db.backup
sqlite3 ~/.claude-memory/memory.db \
  "DELETE FROM memories WHERE created_at < date('now', '-90 days');"

# Rebuild graph
slm build-graph --force

Troubleshooting Commands

Check Installation

# Verify files exist
ls -la ~/.claude-memory/
ls -la ~/.claude-memory/memory.db

# Check slm command
which slm

# Test database
sqlite3 ~/.claude-memory/memory.db "SELECT COUNT(*) FROM memories;"

Fix Permissions

# Fix directory permissions
chmod 755 ~/.claude-memory/

# Fix database permissions
chmod 644 ~/.claude-memory/memory.db

# Fix script permissions
chmod +x ~/.claude-memory/*.py

Database Repair

# Check integrity
sqlite3 ~/.claude-memory/memory.db "PRAGMA integrity_check;"

# Rebuild indexes
sqlite3 ~/.claude-memory/memory.db "REINDEX;"

# Vacuum
sqlite3 ~/.claude-memory/memory.db "VACUUM;"

Reset Everything

# Backup first!
cp -r ~/.claude-memory/ ~/.claude-memory.backup/

# Soft reset (clear memories, keep structure)
rm ~/.claude-memory/memory.db
cd ~/path/to/SuperLocalMemoryV2
./install.sh

# Hard reset (complete removal)
rm -rf ~/.claude-memory/
cd ~/path/to/SuperLocalMemoryV2
./install.sh

Quick Reference Table

Task Command Notes
Save memory slm remember "text" Basic save
Search slm recall "query" Multi-method search
List recent slm list Last 10 by default
System status slm status Health check
Build graph slm build-graph Improve search
Switch profile slm switch-profile <name> Change context
View patterns python3 ~/.claude-memory/pattern_learner.py list Learned preferences
Database backup cp ~/.claude-memory/memory.db backup.db Safety first
Check integrity sqlite3 ~/.claude-memory/memory.db "PRAGMA integrity_check;" Verify DB
Count memories sqlite3 ~/.claude-memory/memory.db "SELECT COUNT(*) FROM memories;" Total count

Related Pages

Created by Varun Pratap Bhardwaj Solution Architect • SuperLocalMemory

GitHub • Issues • Wiki

SuperLocalMemory V3 — Your AI Finally Remembers You. 100% local. 100% private. 100% free.

Part of Qualixar | Created by Varun Pratap Bhardwaj | GitHub

SuperLocalMemory V3

Getting Started

Reference

Architecture

Enterprise

Release Notes

V2 Documentation

Clone this wiki locally