Learn how to interpret your dashboard metrics, track token savings, and optimize your compression strategy for maximum cost reduction.
Your dashboard displays real-time metrics to help you monitor and optimize your token compression:
The number of tokens in your original prompt before compression. This is what you would normally send to the LLM.
Formula:
Counted using the LLM's tokenizer (e.g., GPT-4 tokenizer)Example:
'Hello, how are you today?' = 7 tokens
The number of tokens eliminated through compression. This directly translates to cost savings.
Formula:
Input Tokens - Compressed Tokens = Tokens SavedExample:
1000 input - 650 compressed = 350 saved
The percentage of tokens reduced from the original prompt. Higher is better.
Formula:
(Tokens Saved / Input Tokens) × 100%Example:
(350 / 1000) × 100% = 35% compression
The cumulative cost savings across all your compression requests, calculated based on LLM pricing.
Formula:
Tokens Saved × Token Price (input + output)Example:
10M saved tokens × $0.00003 = $300 saved
LessToken uses proprietary intelligent compression algorithms to automatically reduce your token count while preserving semantic meaning and context.
Our system analyzes your content and applies the optimal compression strategy automatically - no configuration or technical knowledge required.
Average Token Reduction
For Structured Content
Best Case Scenarios
Why it matters: With an average of 20-25% compression, processing 1 million tokens per month saves you 200,000-250,000 tokens, translating to $6-$7.50 saved monthly at GPT-4 pricing.
Visualizes your token consumption over time, showing both input and saved tokens. Look for:
Shows your average compression ratio across all requests:
20-35%
Typical for mixed content
35-50%
Optimized JSON and code
50%+
Highly formatted content
Translates token savings into dollar amounts based on current LLM pricing:
Example Calculation (GPT-4):
Well-formatted JSON and code compress better than plain text. Use structured data when possible.
Avoid repeating instructions or context. State things once clearly instead of multiple times.
If compression takes too long, adjust maxTimeMs option to balance speed vs compression ratio.
{
"endpoint": "/api/users",
"method": "GET",
"description": "Fetch all users from database",
"parameters": {
"page": "number",
"limit": "number"
},
"response": {
"status": 200,
"data": "Array of user objects"
}
}{"endpoint":"/api/users","method":"GET","description":"Fetch all users from database","parameters":{"page":"number","limit":"number"},"response":{"status":200,"data":"Array of user objects"}}Impact at Scale
If you send 1,000 similar prompts per day, this saves 29,000 tokens daily (870,000 monthly), translating to $26+ saved per month at GPT-4 pricing.
Review your dashboard daily to catch anomalies early
Aim for consistent 30%+ compression on formatted content
Monitor month-over-month savings growth
Now that you understand the metrics, explore advanced features and integrations: