Meetily v0.3.0: Import Audio Files, Retranscribe, and What's Coming Next

What's New in Meetily v0.3.0?

Meetily v0.3.0 is the biggest Community Edition update since we hit 10,000 GitHub stars. The focus: making Meetily useful beyond live meetings.

Until now, Meetily only transcribed meetings in real time. If you had a recording from last week, a podcast episode, or an interview you recorded on your phone, you were out of luck. That changes today.

Here's everything in v0.3.0:

• Audio file import - Drag and drop any audio file for local transcription
• Retranscription - Re-run transcription on existing recordings with a different model or settings
• Improved transcription quality - Parallel resampling handles larger files without degradation
• Beta feature gating - New system to preview experimental features before they're fully stable
• Memory optimization - Transcription engine unloads after batch jobs to free system resources

How Does Audio Import Work in Meetily?

Audio import is straightforward. Open Meetily, click "Import Audio," and select your file. Meetily processes it locally using Whisper or Parakeet transcription engines - the same ones used for live meeting transcription. The entire process happens on your device. No upload, no cloud, no third-party processing.

Supported use cases:

• Recorded interviews - Import audio from field recorders or phone recordings
• Podcast episodes - Get searchable transcripts of any podcast locally
• Lecture recordings - Students and researchers can transcribe class recordings
• Old meeting recordings - Transcribe recordings from before you had Meetily
• Voice memos - Turn voice notes into structured text with AI summaries

The feature works with common audio formats and supports 99+ languages, same as live transcription.

What Is Retranscription and Why Does It Matter?

Retranscription lets you re-process an existing recording with different settings. Say you transcribed a meeting with the tiny Whisper model because you needed speed, but the transcript has some errors. Now you can retranscribe that same recording with the large model for higher accuracy - without re-recording anything.

This is useful when:

• You want to try a different transcription model on the same audio
• Audio quality was poor and you want to experiment with settings
• You initially used a faster, smaller model and now want more accuracy
• A new model version becomes available and you want to compare

Both audio import and retranscription are shipping as beta features. We built a new feature gating system to let early adopters try experimental features while we stabilize them based on real-world feedback.

What Improved in Transcription Quality?

This release includes a rewrite of the audio resampling pipeline. Previously, large audio files could hit edge cases during sample rate conversion, leading to degraded transcription quality. The new parallel resampling approach handles files more reliably and processes them faster on multi-core machines.

This is an under-the-hood improvement. You won't see a new button or menu item, but your transcripts should be noticeably better for longer recordings.

We also fixed a memory leak where the transcription engine stayed loaded in RAM after batch jobs completed. Meetily now properly unloads the engine after processing, freeing up system resources for your other work.

What Is Coming Next for Meetily?

We're actively building four features for the next releases. Consider this a preview of what's on the roadmap.

Speaker Diarization - Know Who Said What

This is one of our most requested features. Speaker diarization identifies different speakers in a conversation and labels each segment of the transcript accordingly. We have a proof of concept working internally, and it's a priority for the next major release.

For meeting transcription, this changes everything. Instead of a wall of text, you'll see clearly attributed dialogue - useful for meeting minutes, interview transcripts, and any multi-speaker recording.

Multi-Language Summary Generation

Today, Meetily transcribes in 99+ languages. But summaries are English-only. We're building support for generating meeting summaries in the same language as the transcript, or in a different language entirely.

For multilingual teams, this means you can transcribe a meeting in German and generate the summary in English for the wider team, or vice versa. All processed locally if you're using local AI models.

Inline Transcript Editing

Sometimes the AI gets a word wrong. Currently, there's no way to fix it without exporting and editing externally. Inline editing will let you correct mistakes directly in the app - fix a misheard name, adjust formatting, or clean up technical terms without leaving Meetily.

Dark Mode

Late-night meetings deserve a UI that doesn't burn your retinas. Dark mode is coming across the entire interface.

How to Update to Meetily v0.3.0

Updating is simple:

1. Download the latest installer from meetily.ai/downloads or GitHub releases(opens in new tab)
2. Run the installer - it will update your existing installation
3. macOS: Download the .dmg file and drag to Applications (replaces the old version)
4. Windows: Run the .exe installer and follow the prompts

Your existing meeting transcripts and settings are preserved during the update.

Why Build Audio Import Into a Meeting App?

You might wonder why a meeting transcription tool needs file import. The answer came from our users.

Researchers told us they record interviews in the field and need transcripts later. Podcasters wanted searchable archives of their episodes. Journalists wanted to process source recordings offline. Sales teams had call recordings from before they started using Meetily.

All of them wanted the same thing: a way to get high accuracy transcripts from existing audio files without uploading to a cloud service. Meetily already had the transcription engine running locally. Audio import just gives it a new input source.

We let our users guide the product. Every feature in v0.3.0 was built based on real feedback from people actually using Meetily.

Meetily is built by Zackriya Solutions(opens in new tab). The source code is available on GitHub(opens in new tab).