llmcache

Cache LLM responses. Save tokens, save money.

Installation

npm install -g pnkd-llmcache

Quick Start

# Initialize cache
llmcache init

# Cache a response
llmcache set "What is AI?" "AI is artificial intelligence..."

# Get cached response
llmcache get "What is AI?"

# View statistics
llmcache stats

Commands

FREE Commands

| Command | Description | |---------|-------------| | init | Initialize a new cache | | set <prompt> <response> | Cache a prompt/response pair | | get <prompt> | Get cached response | | list | List cached entries | | stats | Show cache statistics | | clear | Clear cache entries | | search <query> | Search cached prompts | | export [file] | Export cache to JSON | | import <file> | Import cache from JSON |

PRO Commands

| Command | Description | |---------|-------------| | cost | Show cost savings report | | similar <prompt> | Find similar cached prompts | | sync <action> | Sync cache between machines | | serve | Start HTTP server |

Options

Global Options

• -g, --global - Use global cache (~/.llmcache/cache)
• -p, --path <path> - Custom cache path

Set Options

• -m, --model <name> - Model name (default: "default")
• -t, --ttl <duration> - Time to live (PRO): 7d, 24h, 30m
• --tags <tags> - Comma-separated tags (PRO)

Get Options

• -m, --model <name> - Model name (default: "default")
• -r, --raw - Output only the response
• -o, --output <file> - Write response to file

Using Files

Use @ prefix to read from files:

llmcache set @prompt.txt @response.txt
llmcache get @prompt.txt

FREE vs PRO

| Feature | FREE | PRO | |---------|------|-----| | Cache entries | 50 max | Unlimited | | Response size | 10KB | 10MB | | Storage backend | JSON | JSON, SQLite, Redis | | Cost tracking | - | Yes | | Semantic search | - | Yes | | Compression | - | Yes | | HTTP server | - | Yes | | Team sync | - | Yes | | TTL expiration | - | Yes |

PRO License

$18.99 one-time payment

Purchase at: https://pnkd.dev/llmcache

# Activate license
llmcache license activate LMC-XXXX-XXXX-XXXX-XXXX

# Check status
llmcache license status

Programmatic API

const llmcache = require('pnkd-llmcache');

// Initialize
llmcache.init();

// Cache response
llmcache.set('What is AI?', 'AI is...', 'gpt-5-turbo');

// Get cached response
const result = llmcache.get('What is AI?', 'gpt-5-turbo');
if (result) {
  console.log(result.response);
}

// Get stats
const stats = llmcache.stats();
console.log(`Entries: ${stats.entries}`);

HTTP Server (PRO)

llmcache serve --port 3377

Endpoints

• GET /health - Health check
• GET /cache?prompt=...&model=... - Get cached response
• POST /cache - Set cache entry (JSON body)
• GET /cache/list - List entries
• GET /cache/search?q=... - Search entries
• GET /stats - Get statistics

More PRO Tools

• ctxstuff PRO - Pack code for LLMs ($14.99)
• aiproxy PRO - One API for all LLMs ($18.99)

Support

• Issues: https://github.com/pnkd-dev/llmcache/issues
• Website: https://pnkd.dev/llmcache

License

MIT - pnkd.dev