How to bulk export all your Otter.ai transcripts

Last updated 4 October 2026 · 9 min read

Otter is good at turning a meeting into text and bad at giving that text back. If you have more than a handful of conversations you want to keep, you have almost certainly already discovered that there is no “export everything” button anywhere in the product.

What Otter actually allows

This part is not opinion — it is Otter's own documentation. Their help centre states that bulk export to Google Drive “isn't supported at this time”, and that bulk exporting multiple conversations is available on Business and higher plans.

Method 1 — Export one at a time

This is the officially supported free route. Open a conversation, click the export or download control, pick a format, save the file. Repeat.

The real cost is not the clicking — it is these three things:

• It doesn't scale. Twenty conversations is an annoying afternoon. Two hundred is a project you will not finish.
• It is lossy in the ways that matter. Otter's own export gives you the transcript text, but speaker labels, timestamps and action-item structure survive inconsistently depending on which format you pick.
• It competes with your own attention. Every conversation is a decision — where did I save the last one, which format, did I already do this one?

If you have fewer than about fifteen conversations worth keeping, do this and stop reading. Genuinely. It is not worth automating a twenty-minute job.

Method 2 — Pay for one month of Business

A single month on the Business tier unlocks the multi-select export. You export what you need, then downgrade or cancel.

Three things to plan for before you do this:

• Pricing is per seat, and Business is sold in seats. If your workspace has five people on it, one month of Business may cost several times what you assumed.
• Annual billing is usually what's advertised. The headline per-month price is often the annual-commitment figure. Monthly is typically higher.
• You still don't get one archive. You get a batch of individual downloads, which you then have to organise yourself.

The honest case for this method: it is officially supported, it works, and it is the least likely to break. If you have a corporate card and one afternoon, this is a perfectly reasonable answer and you can skip the rest of this page.

Method 3 — The Google Drive route (it doesn't work)

A lot of older blog posts and forum answers suggest wiring Otter to Google Drive for a bulk pull. That advice is out of date. Otter's help centre is explicit that bulk exports to Google Drive are not supported.

What does exist: Otter can send individual conversations to a connected Drive account, and there are automation platforms that will relay conversations one by one. Both end up back at Method 1 with extra steps — and the automation route typically requires a paid automation subscription on top of your Otter plan.

Method 4 — Script it yourself

If you are comfortable with browser devtools, this is the route that actually gives you a complete, structured archive. The shape of it:

1. Open Otter's conversation list while signed in

   In devtools, watch the Network tab while you scroll. Note the request that returns the conversation list, and the request that returns a single conversation's content.

2. Confirm both are reachable with your existing session

   If the list endpoint returns JSON with your conversations in it, you have what you need. If it returns a login wall, the scriptable route is closed to you.

3. Paginate until exhausted

   These endpoints are almost always paginated. Loop until you have as many items as the UI claims you have. This is the step people get wrong — a script that silently stops after page three looks like it worked.

4. Fetch content, then write files locally

   One file per conversation, plus an index. Do the writing in the browser rather than sending data anywhere.

Two honest warnings. First, a home-made script against a web app breaks whenever the front-end changes — you will be maintaining it. Second, you have to be careful how aggressively you hammer the endpoints. Pulling a few hundred conversations as fast as the network allows is the sort of traffic pattern that gets an account flagged, and “I wrote a scraper” is not a defence you want to be having with support.

If you would rather not write it yourself, there are ready-made open-source tools that do exactly this — several, and some are good. Worth knowing before you pick one: most of them authenticate by taking your Otter email and password rather than reusing the session your browser already has, they need Python and a terminal to run, and their own READMEs warn that they break when Otter changes its internals. That is a perfectly reasonable trade for a developer. It is a wall for anyone else.

Method 5 — Use an export extension

A browser extension is Method 4 with the devtools part solved and the rate-limiting part already thought through. It runs the same read-only requests your own session already makes, converts the responses into files, and hands them to your browser's download system.

What to check before you install any of them — this applies to ours as much as anyone's:

• Does it have a server? If the extension uploads your transcripts anywhere to process them, that is a much larger commitment than the problem requires. It should be able to do the whole job locally.
• Is it read-only? It should never edit, delete or share anything in your account. Only read.
• What permissions does it ask for? It should ask for the one site it operates on. An extension that wants tabs, cookies, your history, or <all_urls> is asking for far more than an export tool needs.

Which format should you actually pick?

Think about the next ten years, not about this afternoon. The export you cannot read in 2036 is not an export.

The pragmatic answer is both: Markdown so a human can read it in ten years, JSON so a program can process it tomorrow. They are cheap to produce together and expensive to reconstruct apart.

How to check your export is complete

The single most common way to lose data during a migration is to believe the export finished. Before you cancel anything, confirm these three things:

1. Count. The number of files you exported should match the count Otter's UI shows for your conversation list. Not approximately — exactly.
2. Check the date range. Open your oldest file and your newest. If the oldest one is from six months ago and you know you have notes from three years ago, your conversation list was paginated and your exporter stopped early.
3. Open three files at random and read the middle. Not the first line — the middle. Truncated transcripts almost always fail in the middle, where a chunked fetch gave up.

Do this while you still have access. Re-exporting is trivial if you catch it now and impossible if you catch it after you downgrade.

FAQ