AutoSubs 3.10.1

AutoSubs turns the speech in your video into ready-made subtitles, right on your own computer, with no uploads and no subscription.

It works inside DaVinci Resolve, Premiere Pro and After Effects, or as a standalone app that exports SRT files from any audio or video.

Version 3.10.1 is a big feature update, adding new speech models for rare languages, caption style previews and a long list of timing fixes.

What AutoSubs Actually Does

You point AutoSubs at a clip, pick a language and a speech model, and it writes out what was said with the timing attached. It also works out who is speaking and gives each person their own caption colour.

Everything runs locally. The AI models download once, and after that no audio leaves your machine - handy for interviews, client footage or anything under NDA.

The app is free and open source under the MIT licence. One catch for paid work: the optional forced-alignment add-on, which tightens word timing, uses a noncommercial licence.

Where AutoSubs Fits in a Subtitling Workflow

Captioning used to mean a transcription service, a subtitle editor, a sync tool and the video editor. AutoSubs replaces the first step and hands the rest over in formats your other tools already read.

• Record or import. Capture screen recordings with OBS Studio, or clean up noisy audio first in Audacity.
• Trim if needed. Cut out just the segment you need with LosslessCut, which doesn't re-encode.
• Transcribe. Run AutoSubs from Resolve, Premiere Pro, After Effects or the standalone app.
• Fix and sync. Polish the text in AutoSubs' own editor, or open the SRT in a dedicated editor for heavier work.
• Deliver. Keep the captions on the timeline, mux the SRT into a video file, or burn it into the picture.

What's New in AutoSubs 3.10.1

The headline is choice of speech models. Omni-ASR covers more than 1,600 languages, so it's the one to try when your language isn't well served elsewhere.

The new Orukeet model is built on NVIDIA's Parakeet but runs faster and scores better. It is now the recommended pick for European languages.

GigaAM v3 arrives for Russian, with a multilingual GigaAM for Central Asian languages. The interface itself now comes in a full Russian translation too.

• Caption preset previews. See a style before you apply it, in a gallery with thumbnails and your own presets listed first.
• Batch restyling in Resolve. Change the look of captions already on the timeline without regenerating them.
• Text wrap for animated captions in Resolve, plus styling edited directly in Fusion.
• A redesigned processing screen with live progress and several model downloads running at once.
• Better speech detection that no longer splits words at the edges of audio chunks.

The bug fixes matter just as much for long projects. Transcription no longer hangs on long audio files, and interrupted model downloads now resume properly.

Caption timing drift, marker and highlight errors and fractional frame rounding are fixed. So are negative timecodes in SRT exports and captions landing at the wrong offset in Premiere Pro.

On Windows, ONNX models no longer wrongly default to DirectML. On Intel Macs, the Whisper crashes are gone, as are preview crashes on Resolve 21.

Using AutoSubs With DaVinci Resolve and Premiere Pro

In Resolve, open Workspace > Scripts > AutoSubs. Captions land on the timeline as styled, editable caption clips, one colour per speaker.

In Premiere Pro or After Effects, launch AutoSubs first and its extension loads when the Adobe app opens. Premiere Pro gets a caption track, while After Effects turns each subtitle into a text layer.

Pairing AutoSubs With Other Free Tools

For heavy corrections - line breaks, punctuation across hundreds of lines, spellcheck - Subtitle Edit is the usual choice. Tero Subtitler is a lighter option, and both open AutoSubs SRT files as they are.

If captions drift steadily out of step on a long recording, Time Adjuster shifts the whole file in one go. Subtitle Workshop handles finer, line-by-line timing.

Not editing in Resolve? The standalone app pairs well with Shotcut and Kdenlive, which both import SRT onto a subtitle track.

CapCut users can import the SRT for more control than CapCut's built-in auto-captions give.

To keep subtitles switchable, add the SRT to an MKV file with MKVToolNix. To burn them permanently into the picture for social media, use HandBrake.

Need VTT or ASS for an upload platform? The online subtitle converter at convertico.com handles the common formats without installing anything.

System Requirements and Installation

• Windows: 64-bit installer, AutoSubs-windows-x86_64.exe.
• macOS: separate installers for Apple Silicon (AutoSubs-Mac-ARM.pkg) and Intel Macs (AutoSubs-Mac-Intel.pkg) - check Apple menu, About This Mac if unsure.
• Linux: builds for Debian, Ubuntu, Fedora and openSUSE are on the AutoSubs GitHub releases page.
• Disk space for models. Each speech model downloads on first use, and the model picker shows which ones you can delete to free space.

Speed depends on your hardware and the model you choose. A graphics card helps a lot on long files, and on Windows AutoSubs can use Vulkan or DirectML to tap into it.

Who Should Look Elsewhere

If you only want subtitles for a film you are watching, AutoSubs is overkill. VLC media player loads existing subtitle files, and many films already have them.

If you are captioning commercial client work and want the tightest word timing, check the noncommercial licence on the forced-alignment add-on first. The main transcription models work without it.

For manual subtitle tools, extractors and converters, browse the Subtitle Tools category.

Quick questions

Download AutoSubs 3.10.1 and see how it handles your footage, or tell us how you got on if you have used it before.