X Spaces (formerly Twitter Spaces) is the platform's live audio room feature. Founders host AMAs, journalists run roundtables, communities hold weekly conversations. The recordings often hold the most direct, candid commentary from the host you'll find anywhere, but they're locked inside an audio player with no text. This guide shows the cleanest way to turn a saved Space into a searchable transcript with speaker labels, an AI summary and exportable subtitles.

Why X Spaces Are Worth Transcribing

Spaces is where the unfiltered conversation lives. People say things in a live audio room that they'd never put in a polished tweet. Founders break news, investors share theses, creators host long Q&As. Two-hour Spaces happen daily, and they remain undiscovered for most of the audience because audio isn't searchable on Google, on X, or anywhere else.

Once you have the transcript, the Space becomes a research document. You can quote specific lines, search for moments where a particular topic came up, or summarize the whole thing for someone who didn't have 2 hours to listen live. For journalists, analysts and competitive intelligence teams, this is a major time multiplier.

Worth noting: X added an official AI-generated transcript feature for some Spaces, but coverage is inconsistent, the text is buried inside the X player, and there's no export. A standalone transcript fixes all three issues.

Finding a Saved Space Recording

A Space is only transcribable if the host saved a recording. Here's how to check:

  1. Open the host's X profile.
  2. Look for a Space recording card pinned at the top of their feed, or scroll through their recent tweets, replays usually appear as a tweet with an audio player attached.
  3. Click the Space recording. You should see a player with playback controls.
  4. Copy the URL from your browser address bar. It will look like x.com/i/spaces/<ID> or x.com/i/broadcasts/<ID>.

If no recording exists, you can't transcribe after the fact, unless someone recorded the audio themselves and is willing to share the MP3 file.

How to Get the Text Step by Step

1

Copy the Space URL or download the recording

First try: copy the Space recording URL and paste it into Dokitscript. If the URL is rejected because the Space was archived in an unusual way, the fallback is to download the audio (some browser extensions and recording tools can capture the playback as an MP3), then upload that file via the Upload tab.

2

Open Dokitscript and paste or upload

Go to dokitscript.com. Paste the Space URL into the input on the home page, or click Upload and attach the MP3/M4A audio file. For Spaces longer than a few minutes, you'll want to be on a Pro or Business plan so the duration cap doesn't truncate the file.

3

Set language and transcribe

Auto-detect handles most cases, but if your Space mixes languages or uses heavy jargon, set the language manually for better accuracy. Hit Transcribe. A 1-hour Space takes 5 to 8 minutes of processing time on average.

4

Review, summarize, export

Once the transcript appears, run Summary for a top-line view, then explore the full text. Use Ctrl+F to find specific topics. Export TXT, DOCX or SRT. Save the transcript to your History page for later reuse.

Turn a 2-hour Space into searchable text

Full transcript, speaker labels, AI summary, all in one place.

Transcribe a Space now โ†’

Speaker Detection for Multi-Voice Spaces

The Business plan unlocks automatic speaker detection. Each line in the transcript is labeled with which speaker was talking:

[00:00] Speaker 1: Welcome everyone, thanks for joining...
[00:42] Speaker 2: Glad to be here, can you walk us through the launch?
[01:15] Speaker 1: Sure, so the way we approached it was...
[03:22] Speaker 3: Quick question on pricing...

You can rename "Speaker 1" to the host's actual name (e.g. "Alex"), and that label is applied across the whole transcript wherever Speaker 1 spoke. Big quality-of-life win for writing show notes, interview articles, panel recaps, executive briefings.

Handling Long Spaces (1 to 3 Hours)

Many founder Spaces run 90 minutes to 3 hours. Here's the approach by duration:

Under 35 minutes

Common for shorter Spaces. Pro plan ($14.99/month) covers this without splitting. Unlimited monthly transcriptions on Pro (1,300 minutes fair-use), useful if you process a lot of Spaces.

35 minutes to 5 hours

Business plan ($79.99/month) is built for this length. 5-hour cap plus speaker detection. The Space's full transcript fits in a single job, with proper attribution.

Over 5 hours

Split the audio into segments under the 5-hour cap. Most audio editors (Audacity is free) make this trivial. Transcribe each segment separately, then combine the text files. Even with the split overhead, this is faster than hand transcription by a factor of 30.

Summarize the Whole Space in One Paragraph

Right under the transcript, you'll find AI buttons. Two of them are especially useful for long Spaces:

๐Ÿ“„

Summary

One paragraph capturing the main thread of the conversation. Perfect for newsletter recaps or "here's what you missed" posts.

๐Ÿ“‹

Key Points

Bulleted list of the major takeaways, usually 5 to 8 bullets. Great for memos, briefings, internal team updates.

๐Ÿ’ฌ

Ask a question

Type any question and get an AI answer drawn from the transcript with real-time web sources. "Did the founder mention pricing?" "What's the launch date?" Get answers in seconds.

๐Ÿ“

Blog Post

Turn the transcript into a structured article with headings and intro/outro. Useful for converting a Space into a long-form post for your blog or Substack.

Frequently Asked Questions

If the host saved the Space as a public recording on X, paste the recording URL directly. If the Space wasn't saved, you'd need a screen or audio recording made during the live session, then upload that file. Most large Spaces are saved by default but check the host's profile first.
Yes, especially on the Business plan which includes speaker detection. Each speaker is automatically labeled (Speaker 1, Speaker 2, etc.) and you can rename labels to actual names once for the whole transcript. Free, Starter and Pro return a single-speaker transcript without labels.
Spaces commonly run 30 minutes to 3 hours. Dokitscript caps clip length per plan: Free 3 minutes, Starter 8 minutes, Pro 35 minutes, Business 5 hours. Pro covers short Spaces, Business covers most long-form conversations. Anything over 5 hours can be split into 2 or 3 segments and transcribed separately.
Yes. The transcript captures everyone who spoke during the recording. With speaker detection on the Business plan, each line is labeled with which speaker was talking, so you can attribute quotes correctly even in a 5-speaker conversation.
Yes. Export SRT to use the Space transcript as captions in a video version of the conversation. Some hosts re-edit their Spaces into YouTube uploads with a static graphic, the SRT file works perfectly for adding subtitles to that video.
Yes. After transcription, click Summary or Key Points. The AI condenses a multi-hour conversation into a paragraph or a bulleted list of the main topics covered. Perfect for sharing a recap with people who couldn't join live, or for archiving important takeaways in your notes.
Yes. Click Translate after the transcript is ready and pick any of 90+ supported languages. Useful for global Spaces or for sharing English Spaces with non-English speaking team members.

Make every X Space readable, searchable, shareable

Full transcript, AI summary, exports. Free plan to try, no credit card.

Get started free โ†’

Related: Transcribe a Twitter/X video ยท Twitter thread to text ยท Transcribe a podcast ยท Transcribe a Zoom meeting ยท Create podcast show notes ยท Generate meeting minutes