hama on-device NLP

Think Outside the Cloud

On-device language processing in Python, Node, and the browser

Browser demo Browse the docs

Targets

Python · TypeScript

Browser and server runtimes.

Latency

≈ 7ms / batch

Measured locally on Apple Silicon.

uv pip install hama
from hama import G2PModel

model = G2PModel()
result = model.predict(
    "I saw her duck.",
    preserve_literals="punct",
)
print(result.ipa)
print(result.display_ipa)

Cheap. Fast. Accurate. Pick three.

hama makes local natural language processing an easy choice.

Tasks

Text → IPA

Available

IPA → Text

Coming soon

Audio → IPA

Available

Text → Embeddings

Coming soon

Audio → Text

(Speech recognition)

Coming soon

Features

Apache 2.0 license

Code-mixed input

Input / output character mapping

Python, Node, and browser runtimes

Supported languages

IPA reference
English French German Italian Japanese Korean Portuguese Spanish Dutch Thai (Central) Chinese (Simplified) Chinese (Traditional)

Try it out

Try sample inputs or type your own text.

Loading model weights in the browser… (Inference runs locally.)
Input text
Display IPA

—

Alignments

Blog

View all

[Release] hama v1.3.8: G2P asset refresh

3/19/2026

[Release] hama v1.3.7: Browser exports

3/16/2026

[Release] hama v1.3.6: Browser ASR

3/11/2026

[Release] hama v1.3.5: Display IPA

3/11/2026

[Release] hama v1.3.4: Waveform ASR

3/11/2026

Documentation

Browse docs

Getting started

Introduction

Overview, guiding principles, and supported languages.

Getting started

Installation

Set up Python and TypeScript runtimes for development or deployment.

Reference

APIs

Detailed reference for model classes and helper functions.

Reference

IPA Reference

Practical IPA symbol table and interpretation guide for hama outputs.

Help build the next generation of language intelligence.

Sponsorship supports infrastructure, documentation, and model maintenance.

Sponsor on GitHub Sponsor list