TIL: The free plan transcripts are stored for only 7 days. Almost lost important data.

Just had one of those "oh crud" moments I thought I'd share. You know how you get lulled into a nice free tier? Fireflies.ai is fantastic for turning my team's rambling meetings into something searchable. Been using it for months on the free plan to track action items from our weekly syncs.

Well, I went back today to find a specific technical decision from about three weeks ago. Poof. Transcript and recording, gone. That's when I actually read the fine print on the free plan: **transcripts are only stored for 7 days**. I'd just assumed, since it was in my notebook, it was... mine. Nope. The meeting itself wasn't recorded anywhere else, and my own notes were, let's say, optimistic summaries.

It was a crucial detail about a deployment window agreement with the infrastructure team. Spent an hour reconstructing it from fragmented Slack messages and embarrassing myself with a follow-up call.

So, a friendly heads-up for my fellow self-hosters and lab enthusiasts who might be using this:
* The free plan is great for immediate, short-term reference.
* If it's important, you **must** export it. They offer `.txt`, `.docx`, and `.srt` exports. I've now set a calendar reminder every Friday to export anything valuable from that week.
* For my personal/home lab meetings (yes, I have family meetings about server upgrades 😄), I've actually written a quick script that uses their API to fetch and archive transcripts to my local NAS. Because if you don't control the backup, you don't control the data.

Here's the simple bash loop I use in a scheduled cron job now, in case it helps anyone else:

```bash
#!/bin/bash
# Requires `jq` and your Fireflies API key
API_KEY="your_key_here"
SAVE_DIR="/nas/meeting_archive"

# Get recent transcript IDs (adjust date range as needed)
curl -s -X GET "https://api.fireflies.ai/v1/transcripts?from_date=$(date -d '7 days ago' +%Y-%m-%d)"
-H "Authorization: Bearer $API_KEY"
-H "Content-Type: application/json" | jq -r '.transcripts[].id' | while read -r id; do
# Skip if already archived
if [ ! -f "$SAVE_DIR/${id}.txt" ]; then
curl -s -X GET "https://api.fireflies.ai/v1/transcripts/${id}"
-H "Authorization: Bearer $API_KEY"
-H "Content-Type: application/json" | jq -r '.transcript' > "$SAVE_DIR/${id}.txt"
echo "Archived $id"
sleep 2 # Be polite to the API
fi
done
```

Moral of the story: Always know your data lifecycle, even for SaaS tools. The cloud giveth, and the cloud, well, quietly deleteth after a week.

-- Dad