AI token counter
Paste text for a fast word-based token estimate and public list-price dollars for common API models — not OpenAI tiktoken accuracy. Everything runs in your browser; your text is not uploaded. For monthly scenarios and CloudyBot caps, use the AI cost calculator. Rates are labeled with a last-verified date inside the tool.
Paste text below. We show a word-based token guess (~1 token ≈ 0.75 English words) and list-price dollars for your paste at public API $/1M — not OpenAI tiktoken accuracy. All math runs in your browser; we never receive your paste. Rates as of .
Heuristic: tokens ≈ ceil(word count ÷ 0.75). Rule-of-thumb B (Latin-heavy): ceil(characters ÷ 4) — shown as “Char÷4” below; not used for $ unless you add an advanced toggle later.
Shortcuts: Ctrl+Enter (Windows/Linux) or Cmd+Enter (Mac) copies estimated tokens.
Counts
Why your bill can be higher
- Hidden system prompts, tool definitions, and JSON the model never shows you.
- Retries, multi-step agents, and long tool results.
- Cached-input discounts (or lack thereof) — see the AI cost calculator for monthly + cache % ideas.
Monthly totals: cloudaxisai.github.io/ai-cost-calculator · cloudybot.ai/pricing for CloudyBot caps.
Optional reply cost
Default is prompt (input) tokens only. Model reply using either fixed output tokens or % of input — never both.
List-price cost for this paste
| Model | Prompt $ | Output $ | Total $ |
|---|
Mini glossary
- Est. tokens — word-split heuristic for English-ish prose; production bills use each vendor’s tokenizer.
- Prompt $ — estimated input tokens × public input $/1M.
- Output $ — estimated reply tokens × public output $/1M (zero if “prompt only”).
- Total $ — prompt + output for this single paste scenario (not a monthly subscription quote).
What we did not model
- CJK / mixed-script tokenization, images/audio, tool-use overhead, batch discounts, enterprise deals, tax, FX.
- Exact context-window fit — API tiers vary.
Sources
You have a token estimate. CloudyBot can run prompts on a schedule, watch logs for drift, and keep work inside published plan caps so variable token bills do not surprise finance.
Try CloudyBot FreeFree plan · No credit card · 60-second setup
How it works
Large language model bills are priced in tokens, but most humans think in words, paragraphs, or whole files. This page splits your paste on whitespace, applies a simple words-to-tokens heuristic tuned for English-ish prose, and multiplies by each vendor's published dollars per million input (and optional output) tokens. That gives a directional answer to "is this blob expensive at list price?" before you wire the same text into production.
We also show characters, lines, UTF-8 bytes, Char÷4, and a quiet-reading time estimate so developers can sanity-check logs or emails alongside the token guess. When Latin ratio is low or punctuation density is high, we show a short inline reminder that CJK and code often diverge from word splits — compare the Char÷4 readout and use a real tokenizer in your stack when you are close to context limits.
For monthly totals, blends between models, cache assumptions, and how those numbers sit next to CloudyBot's published caps, open the AI cost calculator. The companion article AI token counter (2026) walks through why we ship both tools.
Common use cases
- Pre-flight a prompt — see whether a wall of text is in the right ballpark before you burn credits in a playground.
- RAG chunk sizing — compare two chunk drafts side by side with the same model table.
- Support email triage — estimate tokens for a pasted thread before summarization.
- Code file sanity check — use Char÷4 alongside the word heuristic when punctuation dominates.
- Teaching hard caps — pair numbers with what hard billing caps mean for stakeholders.
- Handoff to finance — copy TSV into a spreadsheet, then move to the calculator for monthly math.
You just estimated a paste. CloudyBot can automate the next step.
Example: "Every morning, take new rows from this sheet, build a prompt under our token budget, call the API, and Slack me the summary." That is schedules, memory, and optional browser work — not a one-off counter tab.
Try CloudyBot FreeFree plan · No credit card · 60-second setup
Frequently asked questions
Is the AI token counter free?
Yes. There is no signup. All counting runs in your browser and your paste is not uploaded.
Is this as accurate as tiktoken?
No. Vendors bill with their own tokenizers. This page is intentionally simple so you get a fast gut check.
Do you store my text?
No URL hash is used for your paste. Optional theme preference may be stored in localStorage on this device only.
Where is the source code?
Open source under the MIT License at github.com/CloudAxisAi/ai-token-counter. This cloudybot.ai page hosts the same client-side logic for discoverability.
How do I model monthly API spend?
Use the AI cost calculator for messages per day, tokens per message, blends, and CloudyBot tier readouts.
Related tools and reading
Try CloudyBot on the free plan
30 AI Tasks / month · 2 browser sessions (10 min) · 50 MB storage · 7-day chat history
Start free — no credit card