How to Export YouTube Transcripts to Obsidian
Learn how to export YouTube transcripts to Obsidian in seconds. Turn any video into a structured note with AI summaries, key insights, and YAML frontmatter.
Why Obsidian Users Struggle with YouTube
If you use Obsidian for personal knowledge management, you've probably hit this wall: you watch a great YouTube video, take rough notes, and two weeks later you can't find anything. The notes are disconnected from the rest of your vault, and the video is gone from memory.
The problem is that most Obsidian users have no reliable way to export YouTube transcripts to Obsidian in a structured format. YouTube doesn't provide clean transcripts with summaries, metadata, or YAML frontmatter. You're left copying raw text manually — and that's assuming the video even has captions.
This guide shows you exactly how to export YouTube transcripts to Obsidian automatically, with AI-generated summaries, key points, and frontmatter already formatted for your vault. The whole process takes under a minute. No plugins, no manual copying, no rewatching the same video twice to remember what it said.
According to research on video learning, people forget about 70% of what they watch within 24 hours without a note-taking system. For knowledge workers building a second brain, YouTube is one of the largest untapped sources of structured insight. This workflow fixes that.
---
What You Need to Export a YouTube Transcript to Obsidian
You need two things:
- A YouTube video URL — any public video works
- TranscriptAI — an AI transcription tool that exports directly to Obsidian-formatted Markdown
TranscriptAI takes a YouTube URL, transcribes the audio, runs it through an AI to extract key insights and quotes, and generates a `.md` file with YAML frontmatter ready to drop into your vault.
No plugins required. No YouTube API keys. No complicated setup.
---
Step-by-Step: Export a YouTube Transcript to Obsidian
Step 1: Paste the YouTube URL into TranscriptAI
Go to transcriptai.co and paste the URL of any YouTube video. Hit enter.
The tool fetches the transcript automatically — if the video has native captions, it uses those (fast). If not, it uses Whisper AI to transcribe the audio directly.
This takes anywhere from 5 seconds to 30 seconds depending on video length.
Step 2: Review the AI-Generated Note
Once transcription finishes, TranscriptAI shows you the full structured note:
- Summary — a 3-5 sentence digest of the video
- Key points — bullet-point list of the main ideas
- Key quotes — standout phrases worth capturing verbatim
- Topics — auto-detected tags based on the content
- Full transcript — searchable, punctuated text
You can scan the note in 60 seconds and decide whether the video warrants a deeper read — without rewatching a single frame.
Step 3: Click "Export to Obsidian"
In the top-right of the result panel, click the Export button. Select Obsidian from the dropdown.
TranscriptAI downloads a `.md` file with this structure:
---
title: "Video Title Here"
url: "https://youtube.com/watch?v=..."
date: "2026-03-18"
tags: [topic1, topic2, topic3]
source: YouTube
channel: Channel Name
duration: "14:32"
---
Summary
Three to five sentence summary of the video...
Key Points
- First key insight from the video
- Second key insight
- Third key insight
Key Quotes
> "A standout quote from the video."
Full Transcript
The complete punctuated transcript with timestamps...
Every value is YAML-escaped so it won't break your vault's frontmatter parsing. Long values are properly quoted, special characters handled.
Step 4: Drop the File into Your Obsidian Vault
Move the downloaded `.md` file into your Obsidian vault folder — or any subfolder like `Sources/Videos/`.
Obsidian picks it up immediately. You can:
- Link to it from other notes with `[[Video Title]]`
- Query it with Dataview using the `source: YouTube` or `tags` fields
- Search for any phrase from the transcript using Obsidian's full-text search
- Connect it to concept notes with backlinks
One video. One import. Fully integrated into your knowledge graph.
---
What Makes This Different from Copy-Pasting
You could technically get a rough transcript by clicking "More" → "Transcript" on YouTube directly. But YouTube's native transcript has real problems:
- No punctuation (a wall of lowercase text)
- No structure — no summary, no key points
- No YAML frontmatter for Obsidian
- Timestamps mixed into the text make it unreadable
- No AI analysis — just raw words in sequence
Manual copy-paste typically takes 5-10 minutes per video just to get usable text into your vault, and even then you still need to write the summary yourself. TranscriptAI cuts that to under 60 seconds. The AI summary alone saves you from having to re-read raw transcript just to remember what the video covered.
---
Advanced: Using Block References with Timestamps
One of the most useful Obsidian-specific features when working with transcripts is block references. Since the exported transcript includes timestamps, you can reference a specific moment in a video directly from another note.
For example, if a transcript contains:
[12:34] The key insight here is that context window size directly affects retrieval quality...
You can create a block reference `^12m34s` at that line and pull it into your synthesis note:
![[llm-context-windows-talk#^12m34s]]
This creates a citation-like link back to the source video moment — something you can't do with copy-pasted notes. It's especially useful for researchers and students who need to trace insights back to their original source.
---
Using Dataview with Your Transcript Notes
If you use the Dataview plugin in Obsidian, exported transcript notes become queryable. You can build views like:
All YouTube videos watched this month:
TABLE title, date, channel
FROM "Sources/Videos"
WHERE source = "YouTube"
SORT date DESC
All videos tagged with a specific topic:
LIST
FROM "Sources/Videos"
WHERE contains(tags, "machine-learning")
Videos longer than 30 minutes (prioritize reviewing):
TABLE title, duration, date
FROM "Sources/Videos"
WHERE source = "YouTube"
SORT duration DESC
This turns your video history into an actual database you can mine for research, writing, or periodic review.
---
Connecting Transcripts to Your Existing Notes
The real power of exporting YouTube transcripts to Obsidian comes from linking. Once a transcript note lives in your vault:
- Add `[[backlinks]]` from concept notes to source videos — so when you revisit a concept, you see which videos informed it
- Quote the transcript in writing notes with a proper citation trail
- Use block references to pull specific timestamps into other notes without duplicating content
- Search across transcripts and notes simultaneously in Obsidian's global search
For example, if you're building a note on "retrieval-augmented generation," you can link it to three YouTube videos from researchers you transcribed last month, pull in their key quotes, and have a well-sourced synthesis — all without leaving Obsidian.
---
Which Plans Include Obsidian Export
Obsidian export is available on Starter and Pro plans. Free users can still transcribe videos and read the structured note inside the app — the one-click Obsidian export is a paid feature.
If you're just trying it out, the free plan gives you 3 transcriptions to see what the output looks like before committing. No credit card needed.
---
A Real Workflow Example
Here's how one TranscriptAI user — a researcher who studies AI policy — uses the Obsidian export:
- They bookmark YouTube videos throughout the week as they come up
- On Sunday, they batch-transcribe 5-10 videos using TranscriptAI
- They export each to Obsidian with one click
- Inside Obsidian, they tag each note with the relevant policy areas (`[[AI regulation]]`, `[[data privacy]]`)
- When writing a paper or policy brief, they search the vault and pull directly from transcripts — with block references to the exact video moments
The result: a searchable library of video content that feeds directly into their research workflow. No rewatching, no retyping, no lost insights.
---
Conclusion
Exporting YouTube transcripts to Obsidian used to mean awkward copy-paste jobs or third-party scripts that broke whenever YouTube updated its page structure. Now it's a one-click workflow.
If you're already using Obsidian to manage knowledge, YouTube is the biggest gap most users haven't filled. Articles, PDFs, and books are well-served by existing integrations. Video content — which accounts for a growing share of expert knowledge published online — largely isn't. This workflow closes that gap.
Try it with any YouTube video at transcriptai.co — the first three transcriptions are free, no credit card required. Paste a URL, review the note, and export it to Obsidian in under a minute.
---
Related reading: How to Build a Second Brain from YouTube Videos with AI — How to Transcribe a YouTube Video to Text — AI Transcription for Researchers: Analyze Video Interviews Faster