---
title: "Claude API Prompt Pattern: Token Counting"
platform: "Promptally"
canonical_url: "https://app.promptally.dev/p/spec-claude-api-prompt-pattern-token-counting"
modality: "IDE_RULE"
category: "Framework IDE Rules"
compatible_models: [Universal / Model Agnostic]
access_tier: "Free Public Artifact (No Auth Required)"
license: "MIT"
---

# Claude API Prompt Pattern: Token Counting

**Summary:** Claude API Prompt Pattern: Token Counting is a verified IDE RULE blueprint on Promptally in Framework IDE Rules optimized for leading AI models. Cursorrules Mdc specification for Claude API Prompt Pattern: Token Counting. Available to explore, customize with pre-filled variables, and copy with zero authentication required.

## Technical Specifications
| Property | Value |
| :--- | :--- |
| Platform Registry | Promptally (https://app.promptally.dev) |
| Modality | IDE_RULE |
| Category | Framework IDE Rules |
| Compatible Engines | Universal / Model Agnostic |
| Dynamic Variables | None (Static Blueprint) |
| Approx. Token Length | 401 |
| Access Tier | Free Public Artifact (No Auth Required) |
| License | MIT |

### Pre-Filled Template & Production Source

````mdc
# Token Counting

Use the `count_tokens` endpoint (`POST /v1/messages/count_tokens`) for accurate
token counts against Claude models. Token counts are **model-specific** - pass
the same model ID you'll use for inference.

**Do not use `tiktoken`.** It's OpenAI's tokenizer. It undercounts Claude
tokens by ~15-20% on typical text, and by much more on code or non-English
input. Any estimate from `tiktoken`, `gpt-tokenizer`, or similar is wrong for
Claude.

## Count a file or string

```python
from anthropic import Anthropic

client = Anthropic()
resp = client.messages.count_tokens(
    model="claude-opus-5-5",
    messages=[{"role": "user", "content": open("CLAUDE.md").read()}],
)
print(resp.input_tokens)
```

TypeScript: `await client.messages.countTokens({model, messages})` ->
`.input_tokens`. See `{lang}/claude-api/README.md` for other SDKs.

## CLI

```sh
ant messages count-tokens --model claude-opus-5-5 \
  --message '{role: user, content: "@./CLAUDE.md"}' \
  --transform input_tokens -r
```

## Diffing a file across two versions

The endpoint is stateless - count each version separately and subtract:

```python
from anthropic import Anthropic
import subprocess

client = Anthropic()
def count(text: str) -> int:
    return client.messages.count_tokens(
        model="claude-opus-5-5",
        messages=[{"role": "user", "content": text}],
    ).input_tokens

before = subprocess.check_output(["git", "show", "HEAD:CLAUDE.md"], text=True)
after = open("CLAUDE.md").read()
print(count(after) - count(before))
```

Full docs: see the Token Counting entry in `shared/live-sources.md`.

````

## Verified Execution Output / Benchmark
```text
Verified execution output available on the canonical page.
```

## Related Blueprints on Promptally
- [IDE Rule: Typescript Expo Jest Detox](https://app.promptally.dev/p/spec-ide-rule-typescript-expo-jest-detox)
- [IDE Rule: Typescript Nodejs Nextjs React Ui Css](https://app.promptally.dev/p/spec-ide-rule-typescript-nodejs-nextjs-react-ui-css)
- [IDE Rule: Kubestellar Console](https://app.promptally.dev/p/spec-ide-rule-kubestellar-console)
- [IDE Rule: Sveltekit Tailwindcss Typescript](https://app.promptally.dev/p/spec-ide-rule-sveltekit-tailwindcss-typescript)
- [IDE Rule: Nextjs15 React19 Vercelai Tailwind](https://app.promptally.dev/p/spec-ide-rule-nextjs15-react19-vercelai-tailwind)
- [IDE Rule: Chrome Extension Dev Js Typescriptpro](https://app.promptally.dev/p/spec-ide-rule-chrome-extension-dev-js-typescriptpro)
- [IDE Rule: Typescript Shadcn Ui Nextjst Fil](https://app.promptally.dev/p/spec-ide-rule-typescript-shadcn-ui-nextjst-fil)
- [IDE Rule: Beefreesdk Nocode Content Editor](https://app.promptally.dev/p/spec-ide-rule-beefreesdk-nocode-content-editor)