How to Transcribe a Video to Text Free (2026 Guide)

To transcribe a video to text free, upload the file to an online AI transcription tool, wait for the speech recognition pass to finish, then copy or download the result. A typical tutorial or meeting recording takes seconds, not hours. The two decisions that actually matter are which free tier you land in (they range from 45 to 300 minutes a month) and which export format you pick, because plain text, timestamped text, and SRT are useful for completely different jobs.

This guide covers six ways to get a transcript, what each one costs in 2026, and how to pick the export format before you start instead of redoing the job later.

Transcript, Captions, or Subtitles? Get This Right First

These three words get used interchangeably and they are not the same thing.

A transcript is a text document. It reads top to bottom, it is searchable, and you paste it into a doc, an article, or a prompt.

Captions are that same text cut into short timed chunks that appear on screen. They live in a file format like SRT or VTT, or they get burned into the video pixels.

Subtitles are captions in a different language from the spoken audio.

If what you want is text on screen during playback, you do not want a transcript workflow at all: go straight to a subtitle creator or read our comparison of free caption generators. If you want a document you can edit, quote, and repurpose, keep reading. And if you already have an SRT file and need to open or convert it, that is a separate task covered in what an SRT file is and how to work with one.

Method 1: An Online AI Transcript Tool (Fastest for a Single Video)

This is the right default when you have a file and you want text back in under a minute, with nothing to install.

Vidocu's video transcript extractor takes MP4, MOV, WebM, AVI, and MKV, plus audio files like MP3, WAV, and M4A. You can also paste a URL instead of uploading. It returns three formats from the same pass: timestamped text with [MM:SS] markers, clean plain text with no timestamps, and an SRT subtitle file.

How to do it:

1. Open the transcript extractor and either drop your file in or paste the video URL.
2. Wait for the speech recognition pass. Clear English audio typically lands above 95% accuracy.
3. Choose your output: timestamped, plain text, or SRT.
4. Copy it to your clipboard or download the file.

The file ceiling is 500 MB, which covers most screen recordings and meeting captures. Longer or larger source files need a Vidocu account, and the free plan includes one free tool run, so a single transcript costs nothing. Running tools repeatedly is a Pro feature at $39 a month.

The reason to reach for a full video platform rather than a standalone transcriber is what happens next. The transcript is rarely the deliverable. It is the input to a help article, a blog post, or a set of subtitles, and doing those in one place means you are not re-uploading the same video to three services.

Method 2: YouTube's Built-In Transcript (Videos Already on YouTube)

If the video is already published on YouTube, the transcript is sitting there for free.

Open the video, click the three-dot menu below the player, and choose Show transcript. A scrollable panel opens on the right with timestamped lines. There is a toggle to hide the timestamps, after which you can select all and copy.

For your own uploads, YouTube Studio gives you more: open the video, go to Subtitles, and you can download the auto-generated track as an SRT or VTT file.

The honest limits. Auto-captions have no speaker labels, punctuation is inconsistent, and technical terms and product names get mangled more than a dedicated transcription pass would. It is free and instant, which makes it a fine first draft, but plan on editing.

One rule worth stating plainly: downloading transcripts from videos you do not own is a rights question, not a technical one. Use your own content, or content you have permission to use.

Method 3: Tools You Already Pay For (Word, Teams, Zoom)

If the recording came out of a meeting, check what your existing subscription already does before signing up for anything new.

Microsoft Word has a Transcribe feature (Home tab, dictate dropdown) that accepts uploaded audio and video files and returns a timestamped transcript with speaker separation. It requires a Microsoft 365 subscription and caps you at 300 minutes of uploaded audio per month, a limit that cannot be raised by buying more. Users with a Microsoft 365 Copilot license get 30,000 minutes instead.

Teams and Zoom both transcribe recordings natively when the feature is enabled by an admin, with speaker attribution already attached. For internal meetings this is usually the cleanest path, because the audio never leaves systems you have already cleared.

The catch with all three: they are built around meetings. Feed them a screen recording with a voiceover and you get a serviceable transcript with no sense of what was happening on screen.

Method 4: Dedicated Transcription Services

These are built for volume and for meetings, and they are worth it when transcription is the whole job rather than a step in a larger workflow.

Otter.ai is the default for live meeting capture. The free tier gives 300 monthly transcription minutes. Pro is $16.99 per user per month, or $8.33 billed annually, with 1,200 minutes for imported files. Business is $30 per user monthly, or $19.99 annually.

Notta has the most generous free tier at 120 minutes a month, but with a hard 3-minute cap per recording, which makes it a trial rather than a working free plan. Pro is $97.99 a year ($8.17 a month) for 1,800 minutes and 5-hour files. Business is $199.99 a year for unlimited minutes.

Happy Scribe splits AI and human work. The free tier is a 10-minute trial. Basic is 8.50 euro a month billed annually (17 euro monthly) for 120 minutes, Pro is 19 euro annually (29 euro monthly) for 600 minutes, and Business is 59 euro annually (89 euro monthly) for 6,000 minutes. Human transcription starts at 1.75 euro per minute.

Method 5: Editing Suites, If You Edit by Transcript

Descript built its editor around the transcript: delete a sentence in the text and the corresponding video disappears. If that is how you want to work, the transcript is not a side output, it is the timeline.

The free plan covers 60 minutes a month. Hobbyist is $24 monthly or $16 billed annually, Creator is $35 or $24 annually, and Business is $65 or $50 annually.

This is genuinely a different workflow rather than a better transcriber, and it is a real strength of theirs. It is also a lot of software to install if all you needed was a text file. For teams weighing it against a documentation-first workflow, we compared the tradeoffs in our Descript alternatives breakdown.

Method 6: Human Transcription, When Accuracy Is Contractual

AI transcription in 2026 is good. It is not 100%, and there are contexts where the difference matters: legal proceedings, medical records, accessibility compliance you are signing off on, and anything a regulator might read.

Rev is the reference point here. Human transcription is $1.99 per minute. Their AI tier is free for 45 minutes a month, with Essentials at $25.49 per seat monthly billed yearly ($305.90 a year) for 5,000 AI minutes, and Pro at $47.99 per seat billed yearly ($575.88) for 10,000 verbatim minutes.

At $1.99 a minute, a one-hour recording costs about $119. That is the correct price when a mistranscribed sentence creates liability, and clearly the wrong one when you are turning a product walkthrough into a help article.

What Each Method Costs in 2026

Prices verified in August 2026. Vendors change tiers often, so check before you commit to an annual plan.

Which Export Format Do You Actually Need?

Picking this before you start saves a second pass.

If you are heading toward on-screen text in another language, the transcript is only step one: run it through a subtitle translator rather than translating the document by hand and re-timing it.

How to Get a Cleaner Transcript

Accuracy is mostly decided before the AI ever sees the file.

Fix the audio, not the transcript. Background hum, room echo, and clipping cost you more accuracy than any model choice will win back. Running the file through an audio cleanup pass first is usually worth the extra minute.

Feed it audio, not video, when the video is large. If you are near a file size limit, extract the audio track and transcribe that. Nothing in the transcript comes from the pixels.

Give it your vocabulary. Product names, internal acronyms, and people's names are where transcripts break. Fix them once with find and replace rather than reading the whole thing line by line.

Do not transcribe what you meant to summarize. If you want the gist of a 90-minute webinar rather than every word, a video summary gets you there faster, and we compared the options in our AI video summary generator roundup.

What to Do With the Transcript Once You Have It

Very few people want a transcript for its own sake. It is almost always raw material.

Turn it into a help article. A raw transcript is not documentation: it has filler words, backtracking, and no structure. Transcript to help article does the cleanup and structuring pass, and video to documentation skips the intermediate step entirely by producing the guide with screenshots straight from the recording.

Turn it into a written post. Long recordings usually contain one good article. Video to blog post is built for that conversion.

Turn it into subtitles in other languages. The transcript is already the hard part. AI subtitles handles the timing, and translation covers 65+ languages from the same source.

Turn it into a process document. Recorded a workflow rather than a talk? Video to SOP produces the numbered steps instead of a wall of speech.

FAQ

How do I transcribe a video to text for free?

Upload the video to an online AI transcription tool and download the result. Vidocu's transcript extractor accepts files up to 500 MB and returns timestamped text, plain text, or SRT, and the free plan includes one free tool run. Otter (300 minutes a month), Notta (120 minutes), and Descript (60 minutes) all have free monthly allowances too.

How accurate is AI video transcription in 2026?

Clear English audio typically transcribes above 95% accuracy. Accuracy drops with background noise, overlapping speakers, heavy accents, and specialized vocabulary. Expect to correct product names and acronyms regardless of which tool you use. When accuracy is legally consequential, human transcription at around $1.99 per minute is the honest answer.

Can I get a transcript from a YouTube video?

Yes, for videos you own or have permission to use. Open the video, click the three-dot menu below the player, and choose Show transcript. For your own uploads, YouTube Studio lets you download the auto-captions as SRT or VTT. Auto-captions have no speaker labels and inconsistent punctuation, so they usually need editing.

What is the difference between a transcript and subtitles?

A transcript is a plain text document meant to be read. Subtitles are that text cut into timed cues and displayed on screen, usually translated from the spoken language. Same words, different structure and different file formats. A transcript exports as TXT or DOCX, subtitles as SRT or VTT.

Do I need to upload the whole video to get a transcript?

Not always. Nothing in a transcript comes from the picture, so extracting the audio track first and transcribing that works identically and produces a much smaller file. That is the workaround when a recording is over a tool's size limit.

Can I transcribe a video in a language other than English?

Most tools handle major languages, with accuracy varying by language and audio quality. Vidocu covers 65+ languages for translation of the resulting text and subtitles, so a transcript in one language can become on-screen captions in another without re-recording anything.

The Short Version

For a single file, an online AI transcript tool is the fastest route and the format menu (plain text, timestamped, SRT) is the decision that matters. For meetings, check whether Microsoft 365, Teams, or Zoom already covers you before paying for anything. For high monthly volume, a dedicated service earns its subscription. For anything with legal weight, pay a human.

And if the transcript was only ever a means to an end, whether that is a help article, a set of subtitles, or a written guide, the shortest path is to skip the intermediate text file and generate the finished asset from the video directly.

Try Vidocu for free and turn one upload into transcripts, documentation, subtitles, and translations in a single pass.