🚀 Token Estimation for AI Prompts in Laravel — TOON

Write cleaner prompts. Spend fewer tokens. Build smarter AI apps.

🧠 What Problem Does It Solve?

Every time your Laravel app sends data to an AI model — it costs tokens. More tokens = more cost + less context.
But most developers don’t know how heavy their prompt really is.

That’s why TOON includes a built-in Token Estimator — a lightweight tool to quickly measure your prompt before sending it to an LLM.

✨ What Is TOON?

TOON (Token-Optimized Object Notation) is a compact, human-readable data format built for Laravel. It converts PHP arrays / JSON into an AI-friendly format that reduces token usage and improves context clarity.

And now — it can also estimate the token weight of that data.

🔢 Token Estimation — The Feature You Actually Need

TOON gives you a native, zero-dependency PHP method to estimate token usage:

use Sbsaga\Toon\Facades\Toon;

$prompt = Toon::convert($data);            // Converts JSON/PHP array to TOON format
$stats  = Toon::estimateTokens($prompt);   // Estimates token usage

dd($stats);

Example Output:

{
  "words": 20,
  "chars": 182,
  "tokens_estimate": 19
}

💡 Why This Helps Developers

✔ Spot heavy prompts instantly
✔ Reduce AI costs before making API calls
✔ Keep more context inside model limits
✔ Works locally, in CI, or before production
✔ No external tokenizer — fully PHP native

🧭 Practical Use Cases

Use Case	How TOON Helps
Local development	Measure prompt size while designing AI requests
CI check	Fail build if prompt > 3,000 tokens
Debugging	Compare JSON vs TOON format side-by-side
Optimization	Remove unnecessary data & observe impact
LLM cost control	Save real money over large requests

📌 Important Note

This is an approximate estimation, not a billing-accurate tokenizer.
It is model-agnostic, fast, and perfect for comparison / debugging.

🧪 Try It in 20 Seconds

composer require sbsaga/toon

📌 Ready-to-Use Tags (Paste into Medium / Dev.to)

laravel, php, ai, prompt-engineering, chatgpt, openai, ai-tooling, laravel-dev

🧠 “Compress your prompts, not your ideas.” — TOON helps you talk to AI efficiently.

DEV Community