Resona

Resona

by LapisAI

Launch app
Powered by Alibaba Cloud Model Studio

Indonesian speech-to-text, enterprise-grade.

Transcribe YouTube videos and audio files in Bahasa Indonesia with sentence-level timestamps, automatic cleanup, and AI-generated summaries — in seconds.

Start transcribingSee how it works

No sign-up. Paste a YouTube URL or drop an audio file.

resona.xqinformatics.com
00:31Melalui teknologi dual mode, sistem ini menggabungkan tenaga listrik dan hybrid…
00:44Hasilnya, pengalaman berkendara menjadi lebih nyaman, hemat energi…
00:51Dual mode technology sudah diaplikasikan pada beberapa jenis…

Features

Built for Indonesian content workflows.

Every step of the pipeline — extraction, transcription, post-processing, playback — is tuned for Bahasa Indonesia and code-switched technical content.

Bahasa Indonesia native

Tuned for Indonesian audio, including code-switched ID/EN technical terms like dual mode technology or electric power first.

YouTube or upload

Paste a YouTube link or drop any audio/video file — mp3, wav, m4a, mp4, webm. Resona handles extraction and normalization.

Four output formats

Raw transcript with timestamps, cleaned prose, downloadable SRT subtitles, and an AI-generated bullet-point summary.

Click-to-seek playback

Embedded media player with sentence-level navigation. Click any line in the transcript to jump there — the active sentence highlights as it plays.

Fast

A 2-minute clip is transcribed, cleaned, and summarized in under 35 seconds end-to-end. Real-time factor around 0.3×.

History pane

Past transcripts stay accessible in a left sidebar — no account needed. Uploaded audio re-plays from the browser store across sessions.

How it works

Five steps, all automated.

From submission to ready-to-download transcript in under a minute for a typical clip.

  1. 01

    Submit

    Paste a YouTube URL or upload an audio/video file. Resona accepts mp3, wav, m4a, mp4, webm.

  2. 02

    Extract & normalize

    yt-dlp pulls the source audio; ffmpeg transcodes to 16 kHz mono WAV. Files are uploaded to Alibaba OSS as signed URLs.

  3. 03

    Transcribe

    Fun-ASR-MTL transcribes with sentence-level timestamps and word-level alignment. Language hint id ensures Indonesian-first decoding.

  4. 04

    Cleanup & summarize

    qwen3.6-flash adds punctuation, capitalization, and paragraph breaks. qwen3.6-plus generates a bullet-point summary — in parallel.

  5. 05

    Read, play, export

    Browse the transcript with click-to-seek playback. Download TXT, SRT, or JSON. The job auto-saves to your local history.

Stack

Best-in-class models, end to end.

ASRFun-ASR-MTLAlibaba Cloud Model Studio · 31 languages
Cleanupqwen3.6-flashFast text rewriting
Summaryqwen3.6-plusHigher-quality abstraction
Extractionyt-dlp + ffmpeg16 kHz mono WAV
StorageAlibaba OSSSigned URLs · auto-expiring

Try it on your own audio.

Paste a YouTube link or upload a file. No account required.

Launch Resona