AutoSubs 3.6.2

Drop it into DaVinci Resolve, point it at a clip, and within a minute you have a styled subtitle track on the timeline - speakers labelled, colours assigned, timing locked to the frame.

Since version 3.0, AutoSubs also runs as a standalone application, which means you no longer need Resolve open to caption a podcast, an interview, or a screen recording.

Version 3.6.2, released June 9, 2026, is a reliability-focused update that fixes the translation pipeline, model downloads over VPN, and several Windows-specific installation issues.

Where AutoSubs Fits in a Subtitling Workflow

A complete subtitling pipeline used to involve three or four separate programs: a transcription service, a subtitle editor, a synchronisation tool, and the video editor itself.

AutoSubs collapses the first stage into a single click and outputs files that the rest of your pipeline already understands.

The typical flow looks like this. You record or import footage into DaVinci Resolve, or you simply drag a media file into AutoSubs directly.

AutoSubs runs the transcription locally on your device using its Rust backend, which is roughly three times more memory-efficient than the 2.x generation and noticeably faster on diarization.

Speakers are detected automatically, each gets a colour, and the captions land on the timeline as a proper subtitle track - not burned into the picture, so you can still edit them.

From there, the workflow branches depending on what you need.

Quick polish goes straight in AutoSubs' built-in subtitle editor.

Heavier corrections - reformatting line breaks, fixing punctuation across hundreds of cues, or running a spellcheck - are usually faster in a dedicated tool like Subtitle Edit or Tero Subtitler. Both open SRT files exported from AutoSubs without conversion.

What's New in AutoSubs 3.6.2

The 3.6.2 update concentrates on the parts of the workflow that quietly failed in earlier 3.6.x builds, and the translation stage gets most of the attention.

Translated subtitles now follow the formatting rules of the output language rather than the source language - line length and punctuation conventions differ between, say, English and Japanese, and previously the wrong rule set was applied.

Just as important, translation failures are now reported properly instead of silently falling back to the original text, so you will no longer discover an untranslated section halfway through a delivery review.

Model downloads now work correctly through a VPN or system proxy - a frequent blocker for users on corporate networks who could install AutoSubs but never get the AI models onto the machine.

Windows users with accented or special characters in their username or file path no longer hit path-related errors, and the installer itself handles its installation directory more reliably. An export crash affecting certain projects has been fixed, custom prompts from older versions now migrate cleanly, and a DaVinci Resolve timeline refresh issue has been resolved.

The custom prompt editor has also been streamlined into a single text area, with a warning shown if your prompt may conflict with the selected output language.

Key Features of AutoSubs

The defining feature of the 3.x generation is that AutoSubs no longer needs Resolve to function.

The standalone mode accepts any common audio or video format and exports SRT, ASS, or burned-in captions, which makes it a viable transcription tool for podcasters and writers who never touch a video editor. The Rust rewrite also addressed the biggest complaint about earlier versions - idle memory consumption that crept up on long projects.

Speaker diarization produces clean labels and automatic colour assignment per speaker, useful for interviews and panel content.

The advanced speaker styling panel lets you set per-speaker fill, outline, and border colours, so a two-host podcast can have visually distinct captions without manual styling work.

Timing is accurate against variable frame rate sources and drop-frame timecode, two areas where AI subtitle tools historically produce drift.

Transcription supports many languages, with translation available for non-English source audio - and with 3.6.2, that translation stage is finally dependable enough for client work. Multi-line subtitle support and a resizable desktop viewer round out the creator-focused additions.

Pairing AutoSubs with Other Free Tools

AutoSubs handles the AI transcription stage, but a real workflow involves several tools working together. If you are not using DaVinci Resolve as your editor, AutoSubs in standalone mode pairs cleanly with Shotcut and Kdenlive - both accept SRT imports and let you place subtitles on a dedicated track.

Creators working primarily in CapCut can export from AutoSubs and import the SRT for finer control than CapCut's own auto-captions provide.

For trimming the source video before transcription - something worth doing if you only need captions on a specific segment - LosslessCut cuts and joins clips in seconds without re-encoding, so you do not lose quality before AutoSubs touches the file. After captions are finalised, MKVToolNix is the standard tool for muxing the SRT track into an MKV file alongside the video, useful for archiving or distributing content with selectable subtitles.

If your AI-generated captions drift slightly against the video - a common issue with very long recordings or unusual frame rates - Time Adjuster handles linear sync corrections in a single dialog.

For format conversion between SRT, VTT, ASS, and other subtitle types when uploading to platforms with specific requirements, the online subtitle converter at convertico.com handles the common pairs without an installer.

System Requirements and Installation

AutoSubs 3.6.2 is a 64.1 MB download and runs on Windows 10 and Windows 11. The standalone version installs without requiring DaVinci Resolve, although the Resolve integration is the original use case and remains the most polished workflow.

Transcription happens locally on your machine, which means no audio is uploaded to a cloud service - a meaningful difference if you handle confidential interview material or unreleased content.

Installation is a standard Windows installer, and 3.6.2 specifically improves how it handles its installation path, reducing post-install issues.

On first launch, AutoSubs downloads the AI model files it needs for transcription - now working correctly even behind a VPN or proxy. Clear status badges show at a glance which models are installed, which are downloading, and which can be safely deleted to free up disk space.

For users who want to explore alternative subtitle tools beyond AutoSubs' AI approach, the full Subtitle Tools section of free-codecs.com lists more than thirty free editors, converters, extractors, and synchronisers covering every format and use case.

Why Choose AutoSubs Over Online Auto-Caption Services

The main reasons to use AutoSubs instead of a browser-based service like YouTube's auto-captions or a paid SaaS product come down to control, privacy, and integration. Browser tools force you to upload your video and wait, then download a generic SRT that often needs reformatting.

AutoSubs runs offline, gives you styled tracks on the Resolve timeline directly, and never leaves your local machine. For freelancers under NDA, content creators with unreleased material, or anyone working in a low-bandwidth environment, that distinction matters.

The trade-off is hardware - local AI transcription is faster on machines with a recent CPU and at least 8 GB of RAM. The Rust backend lowers the floor noticeably compared to the 2.x releases, but very large files still benefit from a discrete GPU. For routine YouTube videos, podcast episodes, and short-form social content, modern laptops handle AutoSubs comfortably.

Download AutoSubs 3.6.2 free for Windows from the link above and add AI subtitling to your editing workflow without subscriptions or upload limits.