0 characters · 0 words · 0 lines
A rough estimate from character patterns, not the real tokenizer. Looks like prose. For an exact count, call the API's count_tokens endpoint.
Context window
Claude Code also spends context on its system prompt, tools, CLAUDE.md and MCP tool definitions. Run /context to see the real breakdown.
Cost to send it once as input
- Claude Fable 5.1$0.0000 · cached $0.0000
- Claude Fable 5$0.0000 · cached $0.0000
- Claude Opus 5.5$0.0000 · cached $0.0000
- Claude Opus 5$0.0000 · cached $0.0000
- Claude Sonnet 5.5$0.0000 · cached $0.0000
- Claude Sonnet 5$0.0000 · cached $0.0000
Keeping Claude Code's context lean
- Keep
CLAUDE.mdshort; it loads every session. The CLAUDE.md generator keeps it tight. - Use
/clearbetween unrelated tasks and/compactwhen a long task fills up. - Hand searches to a subagent so only its summary comes back.
- Show context usage all the time with a custom status line.
Turn token counts into a monthly bill with the API cost calculator.
Frequently asked questions
How many tokens is a word for Claude?
For English prose, a word is usually one to two tokens, and code, numbers and non-English text use more tokens per character. Claude 4.7 and later use a newer tokenizer that produces about 30% more tokens for the same text than Sonnet 4.6 and earlier.
Is this an exact Claude token count?
No. Anthropic does not publish a tokenizer that runs in the browser, so this page estimates from character patterns. For exact counts, use the count_tokens endpoint of the Claude API.
How big is Claude's context window?
All current Claude models have a 1 million token context window. Some older models and settings use 200K. In Claude Code, run /context to see how much of the window your conversation, files and tools are using.
Is my text uploaded?
No. Counting happens in your browser and files you load are read locally. Nothing is sent to a server.
Sources: Claude API pricing (tokenizer note). Claude Code Agents is an independent site, not affiliated with Anthropic.