Word Counter & Token Estimator - GPT & Gemini Cost Tool

Word counter & token estimator for GPT and Gemini. Count CJK, estimate tokens, and project API cost with your pricingโ€”private, client-side. Try it today.

Desktop version uses a split view; mobile version stacks panels so your KPIs remain visible.

Words

0

Characters

0

Sentences

0

Paragraphs

0

Input Tokens

0

Output Tokens

0

Reading Time

0.0 min

Speaking Time

0.0 min

Avg Sentence Length

0.0 w/s

Estimated USD Cost

$0.0000

Estimated Local Cost: 0.00

Goal Progress

0.0% achieved (0 / 1,200 words)

All word & token counting is performed 100% locally in your browser. No text is ever uploaded.

How to Use

  1. Paste your article, prompt, or script.
  2. Review live counts and token estimates on the right.
  3. Select a model and enter your pricing, FX rate, and output multiplier.
  4. Click Copy Summary or Export CSV for reporting.

Core Features

  • Count words, characters, sentences, and paragraphs in real time.
  • Estimate input/output tokens for GPT- and Gemini-style models.
  • Calculate projected API cost from the pricing and FX rate you set.
  • Measure CJK character counts for Chinese/Japanese/Korean text.
  • Forecast reading and speaking time from adjustable WPM settings.
  • Export a shareable summaryโ€”everything stays local and private.

Related Tools

Calculation Logic

  • Word splitting can use Intl.Segmenter when available for better accuracy.
  • Token counts are estimates: tokensโ‰ˆcharsratiotokens \approx \frac{chars}{ratio} (ratio varies by model and language).
  • Cost is computed from the prices you input, avoiding stale pricing tables.

FAQ

  • Is it free? Any limits?
    • Yesโ€”core counting and token estimation are free.
  • Is my text private? Do you upload it?
    • No. Everything runs client-side in your browser with no uploads.
  • How do I estimate GPT/Gemini API cost from text?
    • Paste text, choose a model, and enter your input/output prices plus output multiplier.
  • Why can estimated tokens differ from exact billed tokens? (Technical)
    • Tokenization rules differ by model and language; whitespace/punctuation/Unicode can change counts. Estimates are best for budgeting.
  • Does it handle CJK accurately?
    • Yesโ€”CJK-aware counting helps you budget prompts and scripts more reliably.