Your Otter History Has a Limit: How to Export and Keep Every Transcript

The problem is not transcription. It is retention.

Most meeting tools are sold on the live moment: press record, get text, move on. Keeping that text for a year, in a format you control, is treated as an afterthought, and the plan limits reflect it.

Concrete example: Otter Basic.

• Keeps only the 25 most recent conversations
• 300 minutes per month
• Export limited to TXT or MP3

Pass conversation 25 and the oldest one is no longer there. For anyone taking two calls a day, that is less than three working weeks of history. Fireflies has the mirror-image problem on its free tier: no downloads at all without Pro.

If you build your client work on top of these tools — coaching notes, consulting discovery calls, freelance project calls — your archive is quietly rotating out from under you.

This post is a practical fix: export methodically, file transcripts so search actually works, and decide deliberately where the canonical copy lives.

The full guide also lives on the MeetScribe site: Export your Otter transcripts.

Step 1: Export oldest-first

Order matters. Export in reverse chronological order so the conversations closest to the limit are secured first.

1. Sign in to Otter on desktop and open the conversation list.
2. Open the oldest conversation you want to keep.
3. Use the export menu (three dots / share / export, depending on your view).
4. On Basic, export the TXT transcript first. Add the MP3 as well for any call where tone or exact wording might matter later.
5. Rename on save using a stable pattern (see below).
6. Open one exported file and spot-check the first screen of text before batching the rest.
7. Work forward to the newest conversation.

On paid Otter tiers you get more format options. Take the most complete text format available, and still keep audio for high-stakes calls.

Put a recurring monthly slot in your calendar for this. An archive that depends on you remembering is not an archive.

Step 2: Use a naming pattern search can survive

A transcript library fails at retrieval time, not at save time. This pattern is boring and works:

YYYY-MM-DD — Person or client — Short topic.txt
2026-03-14 — Harbour Coaching — Discovery call.txt

Date first gives you chronological sorting for free. Person second matches how you actually remember a meeting ("that call with Harbour, around March"). Topic third breaks ties when you talked to the same person five times.

Avoid version suffixes in transcript names. A transcript is a record of something that already happened — if you export it twice, deduplicate it rather than renaming it final-v2.

Step 3: Decide where the canonical copy lives

Three properties worth optimising for:

Single location. One library, not a spread across the meeting app, your inbox, and a downloads folder. Search only works when the corpus is in one place.

Full text preserved. Summaries and action-item lists are derived views. Keep the underlying transcript so any summary can be checked against the source sentence later.

Local control. Client conversations are business records, often containing pricing, personal context, and commitments. A copy on your own machine cannot be removed by a plan change or a retention policy update. Back it up like any other business record.

What I use for the library step

Disclosure: I built MeetScribe, so take this section as a maker describing their own tool, with its limits stated plainly.

MeetScribe is a desktop app that runs on your own computer:

• Records from the microphone and transcribes locally using faster-whisper. There is a one-time model download of roughly 75 MB; after that, transcription works fully offline.
• No account is required for the core app, and recordings are not uploaded.
• Imports audio, video, and existing transcripts from Fathom, Otter, Fireflies, and Zoom into one searchable library, with deduplication so a repeated import does not create parallel copies.
• Extracts action items with an owner and a due date, and maintains a commitments register — what you owe versus what you are waiting on — keeping the exact source sentence attached to each entry so it can be verified against the transcript.

Limits, stated plainly: it does not automatically add speaker labels on plain recordings, and transcription accuracy depends on audio quality — a quiet speaker on a laptop mic in a noisy room will produce worse text than a close mic in a quiet one, whatever tool you use.

Pricing is $29 one-time for the core app, with the first 5 meetings free. An optional Pro tier at $9/month adds Fathom sync. The core library and import workflow do not need Pro.

You do not need it for the export routine above — a disciplined folder and the naming pattern will carry you a long way. The app exists for the point where the folder stops scaling and you want the imports, search, and commitments list handled in one place.

A minimal monthly routine

1. Export any Otter conversations approaching the 25-conversation edge, oldest first.
2. Rename and file them the same day; do not leave them in downloads.
3. Spot-check one file.
4. Review open commitments from the month's meetings: what you owe, what you are waiting on, and anything past its date.

Twenty minutes a month. The alternative is discovering the limit exists on the day you go looking for the one conversation you actually needed.