# Hebrew & Arabic Audio Transcription + RTL Subtitles (Whisper) (`harushlabs/hebrew-arabic-audio-transcription`) Actor

Transcribe Hebrew, Arabic and 90+ languages from your audio/video files to text, timestamps and RTL-correct SRT/VTT subtitles. Hebrew uses ivrit.ai's Hebrew-tuned Whisper. Runs fully inside Apify - no third-party API key, no scraping.

- **URL**: https://apify.com/harushlabs/hebrew-arabic-audio-transcription.md
- **Developed by:** [Harush Studio](https://apify.com/harushlabs) (community)
- **Stats:** 2 total users, 1 monthly users, 100.0% runs succeeded, 0 bookmarks
- **User rating**: No ratings yet

## Pricing

$80.00 / 1,000 audio minutes

This Actor is paid per event. You are not charged for the Apify platform usage, but only a fixed price for specific events.

Learn more: https://docs.apify.com/actors/running/actors-in-store.md#pay-per-event

## What's an Apify Actor?

An Actor is a serverless cloud program that runs on the Apify platform. It has two run modes.
In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours,
and optionally produces a well-defined JSON output, datasets with results, or files in key-value store.
In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server.

Apify vocabulary and the platform model are defined once, in the agent quickstart at https://apify.com/agents.md.

## How to integrate an Actor?

If asked about integration, you help developers integrate Actors into their projects.
You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready.

Do not guess an integration path. Every one of them is in the agent quickstart at https://apify.com/agents.md: the Apify MCP server, Agent Skills with the Apify CLI, the JavaScript and Python clients, the REST API, and the account-free path for an agent with no human to sign in. It also carries the rule on stating cost before the first paid run.

For examples already wired to this Actor's own input schema, see the [API](#api) section below.

Each client library has reference documentation the quickstart does not restate: [JavaScript/TypeScript](https://docs.apify.com/api/client/js/docs.md) (`npm install apify-client`) and [Python](https://docs.apify.com/api/client/python/docs.md) (`pip install apify-client`).

# README

## Hebrew & Arabic Audio Transcription + RTL Subtitles (Whisper)

Turn **Hebrew** and **Arabic** audio or video into accurate text, timestamped segments and **SRT / WebVTT subtitles that display correctly right-to-left**. Hebrew is transcribed with **ivrit.ai's Hebrew-tuned Whisper large-v3-turbo**, which is much more accurate on Hebrew than generic Whisper; Arabic and 90+ other languages use **Whisper large-v3-turbo**.

Everything runs inside the Actor: **no OpenAI / Deepgram API key, no third-party upload, no scraping**. You provide files you own or are allowed to process; results go only to your own Apify storage.

### What you get

- 📝 **Transcript text** (one line per spoken segment)
- ⏱️ **Timestamped segments** and optional **word-level timestamps**
- 🎬 **SRT and VTT subtitles** with broadcast-style line wrapping (42 chars × 2 lines by default) and the **RTL punctuation fix** (periods, question marks and numbers stay on the correct side in VLC, YouTube, browsers and editors)
- 💾 `.srt` / `.vtt` files saved to the key-value store for direct download
- 🌍 **Auto language detection**, or force Hebrew / Arabic / English / Russian / French / Spanish / Amharic / Yiddish
- 🔁 Optional **translation to English**

### Typical uses

- Subtitles for Hebrew/Arabic courses, webinars, podcasts, marketing videos and Reels you produced
- Meeting and interview transcripts for your team (with your participants' consent)
- Feeding Hebrew/Arabic audio into **RAG, search and AI agents** (LangChain, LlamaIndex, n8n, Make, MCP)

### Input example

```json
{
  "mediaUrls": [{ "url": "https://example.com/my-webinar.mp4" }],
  "language": "he",
  "outputFormats": ["text", "srt"],
  "vocabularyHint": "Apify, ivrit.ai, תל אביב"
}
```

Upload files in Apify Console (they go to your own key-value store) or pass any direct `https://` link: S3, Google Cloud Storage, Dropbox direct links, your own server. Supported: MP3, WAV, M4A, AAC, OGG, OPUS, FLAC, MP4, MOV, MKV, WEBM and anything else ffmpeg can read. **YouTube or social-media page links are not supported.**

### Output example

Real output on a 30-second clip of the Hebrew Spoken Wikipedia article "שחמט" (CC BY-SA 3.0):

```json
{
  "fileName": "he_chess_30s.ogg",
  "status": "ok",
  "language": "he",
  "direction": "rtl",
  "durationSeconds": 30.0,
  "billedMinutes": 1,
  "model": "ivrit-ai/whisper-large-v3-turbo-ct2",
  "text": "שחמט הוא משחק לוח אסטרטגי המיועד לשני שחקנים.\nזהו אחד מהמשחקים השכיחים והמורכבים ביותר הקיימים בתרבות האנושית.\n…",
  "segments": [{ "start": 0.28, "end": 3.32, "text": "שחמט הוא משחק לוח אסטרטגי המיועד לשני שחקנים." }],
  "srtKey": "subtitles-0001-he_chess_30s.srt"
}
```

SRT excerpt:

```
1
00:00:00,280 --> 00:00:03,320
שחמט הוא משחק לוח אסטרטגי
המיועד לשני שחקנים.
```

### Pricing

**$0.08 / minute, $4.80 per hour of audio** (billed per started minute, with a 1-second grace) — Pay per event, you only pay for files transcribed successfully. Failed downloads, unreadable files and silent files are free. A 45-minute Hebrew lecture costs $3.60. Set *Maximum cost per run* and the Actor will stop (or truncate the last file) before exceeding it.

### Tips for best accuracy

- Set `language` explicitly when you know it (faster, avoids detection mistakes on short clips).
- Put names and jargon in `vocabularyHint`.
- Use 8 GB memory (default). More memory = more CPU = faster runs, same price per minute.
- Keep `skipSilence` on: it trims long silences and prevents Whisper from "hallucinating" text in silence.

### Limitations

- No speaker diarization (who said what) in this version.
- Very noisy audio, music or heavy overlapping speech reduce accuracy, as with any ASR.
- Speed (measured): about real time at 8 GB memory (a 3-minute file took ~3.5 minutes plus ~1 minute start-up); 16–32 GB memory gives more CPU cores and proportionally faster runs.

### Privacy & legal

- Process only media you own or have the rights and, where required, the participants' consent to transcribe.
- Files are processed in your own run and are not kept by the developer. Logs contain no URLs, file names or transcript text.
- Models: [ivrit-ai/whisper-large-v3-turbo-ct2](https://huggingface.co/ivrit-ai/whisper-large-v3-turbo-ct2) (Apache-2.0), [Whisper large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) (MIT) via [faster-whisper](https://github.com/SYSTRAN/faster-whisper) (MIT). Media decoding by [FFmpeg](https://ffmpeg.org).

### Related Actors

- **Hebrew & Arabic PDF to Markdown** – RTL-correct document conversion with OCR.
- **RTL Subtitle Fixer & Converter** – fix Hebrew/Arabic punctuation, encoding and timing of existing SRT/VTT files.

# Actor input Schema

## `mediaUrls` (type: `array`):

Direct links to audio or video files you own or are allowed to process (MP3, WAV, M4A, OGG, OPUS, FLAC, MP4, MOV, WEBM...). The Actor does not download from YouTube or social networks.

## `language` (type: `string`):

Hebrew uses ivrit.ai's Hebrew-tuned Whisper model; other languages use Whisper large-v3-turbo. Setting the language explicitly is faster and more accurate than auto-detection.

## `translateToEnglish` (type: `boolean`):

Output an English translation instead of the original-language transcript (Whisper translate task).

## `outputFormats` (type: `array`):

Which outputs to include in each dataset item.

## `wordTimestamps` (type: `boolean`):

Adds per-word timings (slower, ~10-20%). Also improves subtitle splitting.

## `skipSilence` (type: `boolean`):

Voice activity detection removes long silences and reduces hallucinated text.

## `vocabularyHint` (type: `string`):

Optional names, terms or spelling you expect (e.g. product names, people, places). Max 800 characters.

## `subtitleMaxCharsPerLine` (type: `integer`):

Long segments are split into several cues so each line stays within this length (42 is the common broadcast standard).

## `subtitleMaxLines` (type: `integer`):

Maximum lines per subtitle cue.

## `rtlSubtitleMode` (type: `string`):

Keeps punctuation and numbers on the correct side when Hebrew/Arabic subtitles are displayed.

## `saveSubtitleFiles` (type: `boolean`):

Also store subtitle files in the run's key-value store for direct download.

## `maxMinutesPerFile` (type: `integer`):

Longer files are transcribed up to this limit (and flagged as truncated).

## `maxFileSizeMb` (type: `integer`):

Files larger than this are skipped (status: error).

## Actor input object example

```json
{
  "mediaUrls": [
    {
      "url": "https://api.apify.com/v2/key-value-stores/nWOtrXwS4SUzXDleJ/records/sample-hebrew-30s.ogg?signature=are64VSpxvOrzh07O0Ax"
    }
  ],
  "language": "auto",
  "translateToEnglish": false,
  "outputFormats": [
    "text",
    "segments",
    "srt",
    "vtt"
  ],
  "wordTimestamps": false,
  "skipSilence": true,
  "subtitleMaxCharsPerLine": 42,
  "subtitleMaxLines": 2,
  "rtlSubtitleMode": "rlm",
  "saveSubtitleFiles": true,
  "maxMinutesPerFile": 180,
  "maxFileSizeMb": 500
}
```

# Actor output Schema

## `results` (type: `string`):

Dataset with one item per media file: language, duration, text, segments, words, SRT and VTT content, status and errors.

## `subtitleFiles` (type: `string`):

RTL-correct .srt and .vtt subtitle files per transcribed file (keys are also in srtKey/vttKey).

# API

You can run this Actor programmatically using our API. Below are code examples in JavaScript, Python, and CLI, as well as the OpenAPI specification and MCP server setup.

## JavaScript example

```javascript
import { ApifyClient } from 'apify-client';

// Initialize the ApifyClient with your Apify API token
// Replace the '<YOUR_API_TOKEN>' with your token
const client = new ApifyClient({
    token: '<YOUR_API_TOKEN>',
});

// Prepare Actor input
const input = {
    "mediaUrls": [
        {
            "url": "https://api.apify.com/v2/key-value-stores/nWOtrXwS4SUzXDleJ/records/sample-hebrew-30s.ogg?signature=are64VSpxvOrzh07O0Ax"
        }
    ]
};

// Run the Actor and wait for it to finish
const run = await client.actor("harushlabs/hebrew-arabic-audio-transcription").call(input);

// Fetch and print Actor results from the run's dataset (if any)
console.log('Results from dataset');
console.log(`💾 Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`);
const { items } = await client.dataset(run.defaultDatasetId).listItems();
items.forEach((item) => {
    console.dir(item);
});

// 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/js/docs

```

## Python example

```python
from apify_client import ApifyClient

# Initialize the ApifyClient with your Apify API token
# Replace '<YOUR_API_TOKEN>' with your token.
client = ApifyClient("<YOUR_API_TOKEN>")

# Prepare the Actor input
run_input = { "mediaUrls": [{ "url": "https://api.apify.com/v2/key-value-stores/nWOtrXwS4SUzXDleJ/records/sample-hebrew-30s.ogg?signature=are64VSpxvOrzh07O0Ax" }] }

# Run the Actor and wait for it to finish
run = client.actor("harushlabs/hebrew-arabic-audio-transcription").call(run_input=run_input)

# Fetch and print Actor results from the run's dataset (if there are any)
print(f"💾 Check your data here: https://console.apify.com/storage/datasets/{run.default_dataset_id}")
for item in client.dataset(run.default_dataset_id).iterate_items():
    print(item)

# 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/python/docs/quick-start

```

## CLI example

```bash
echo '{
  "mediaUrls": [
    {
      "url": "https://api.apify.com/v2/key-value-stores/nWOtrXwS4SUzXDleJ/records/sample-hebrew-30s.ogg?signature=are64VSpxvOrzh07O0Ax"
    }
  ]
}' |
apify call harushlabs/hebrew-arabic-audio-transcription --silent --output-dataset

```

## MCP server setup

```json
{
    "mcpServers": {
        "apify": {
            "type": "http",
            "url": "https://mcp.apify.com/?tools=fetch-actor-details,harushlabs/hebrew-arabic-audio-transcription"
        }
    }
}
```

The hosted server signs you in with OAuth on first connect, so no API token belongs in this config. Clients without OAuth support can send an `Authorization: Bearer <APIFY_API_TOKEN>` header instead, using a token from API & Integrations in Apify Console (https://console.apify.com/settings/integrations).

## OpenAPI specification

Download the OpenAPI definition: https://api.apify.com/v2/actors/ApnK3AnUnMdn1L3qM/builds/nUmtbVgit4Wp0Otg5/openapi.json