@pr0gramm/fluester

fluester

[ˈflʏstɐ]

Node.js bindings for OpenAI's Whisper. Hard-fork of whisper-node.

Features

• Output transcripts to JSON (also .txt .srt .vtt)
• Optimized for CPU (Including Apple Silicon ARM)
• Timestamp precision to single word

Installation

Requirements

• make
• Everything else listed as required to compile whisper.cpp

1. Add dependency to project

npm install @pr0gramm/fluester

1. Download whisper model of choice

npx @pr0gramm/fluester download

Usage

import whisper from "@pr0gramm/fluester";

const transcript = await whisper("example/sample.wav");

console.log(transcript); // output: [ {start,end,speech} ]

Output (JSON)

[
  {
    "start": "00:00:14.310", // time stamp begin
    "end": "00:00:16.480", // time stamp end
    "speech": "howdy" // transcription
  }
]

Usage with Additional Options

import whisper from "@pr0gramm/fluester";

const filePath = "example/sample.wav", // required

const options = {
  modelName: "tiny.en", // default
  modelPath: "/custom/path/to/model.bin", // use model in a custom directory
  whisperOptions: {
    gen_file_txt: false, // outputs .txt file
    gen_file_subtitle: false, // outputs .srt file
    gen_file_vtt: false, // outputs .vtt file
    timestamp_size: 10, // amount of dialogue per timestamp pair
    word_timestamps: true // timestamp for every word
  }
}

const transcript = await whisper(filePath, options);

Made with

• A lot of love by @ariym at whisper-node
• Whisper OpenAI (using C++ port by: ggerganov)
• ShellJS

Roadmap

• Drop ShellJS, use native/builtin stuff
• Nothing ¯\(ツ)/¯