Skip to main content

​
Installation

You can install bun_nltk using your preferred JavaScript package manager. The package ships with prebuilt native binaries for supported platforms and a WASM runtime for universal compatibility.

​
Install with package managers

npm install bun_nltk

​
Platform support

bun_nltk provides optimized builds for different platforms:

​
Native binaries (fastest)

Prebuilt native binaries are included for:
  • Linux x64: native/prebuilt/linux-x64/bun_nltk.so
  • Windows x64: native/prebuilt/win32-x64/bun_nltk.dll
The library automatically detects your platform and loads the appropriate native binary for maximum performance.
Native binaries provide the best performance, typically 3-10x faster than WASM on most operations.

​
WASM runtime (universal)

For platforms without native binaries, bun_nltk includes a WebAssembly build:
  • All platforms: native/bun_nltk.wasm
The WASM runtime provides:
  • Universal compatibility across all platforms and runtimes
  • Browser environment support
  • Fallback for unsupported architectures (macOS, ARM, etc.)
  • Still faster than pure JavaScript implementations
You can explicitly use the WASM runtime by importing from the WasmNltk wrapper class, even on platforms with native binaries.

​
Verify installation

After installation, verify that bun_nltk is working correctly: 
import { tokenizeAsciiNative, countTokensAscii } from "bun_nltk";

const text = "Hello world! This is a test.";
const tokens = tokenizeAsciiNative(text);
const count = countTokensAscii(text);

console.log(`Tokens: ${tokens.join(", ")}`);
console.log(`Count: ${count}`);
// Output:
// Tokens: hello, world, this, is, a, test
// Count: 6
No build steps or lifecycle scripts are required - bun_nltk works immediately after installation.

​
Environment variables

For advanced use cases, you can customize the native library path: 
export BUN_NLTK_NATIVE_LIB=/path/to/custom/bun_nltk.so
This is useful for:
  • Local development with custom builds
  • Testing new native implementations
  • Deployment environments with specific binary requirements

​
Troubleshooting

​
Platform not supported

If you see an error about missing native libraries: 
native library not found for platform=darwin arch=arm64
Supported prebuilt targets: linux-x64, win32-x64
You have two options:
  1. Use the WASM runtime (recommended):
import { WasmNltk } from "bun_nltk";

const wasm = new WasmNltk();
await wasm.init();

const tokens = wasm.tokenizeAscii("Hello world!");
  1. Build from source (advanced):
# Clone the repository
git clone https://github.com/Seyamalam/bun_nltk.git
cd bun_nltk

# Build native library
bun run build:zig

# Build WASM
bun run build:wasm

​
Import errors

If you encounter TypeScript import errors, ensure your tsconfig.json includes: 
{
  "compilerOptions": {
    "module": "ESNext",
    "moduleResolution": "bundler",
    "types": ["bun-types"]
  }
}

​
Next steps

Quickstart

Learn the basics with a quick example

API reference

Explore all available functions

Build docs developers (and LLMs) love

Get started for freeTalk to us