Zoom Recording Archival

Zoom Recording Archival describes approaches for preserving Zoom meeting recordings – both cloud-hosted and locally saved – into git-annex repositories. Unlike the other tools in this section, there is no single turnkey tool for Zoom archival. Instead, this page documents the building blocks and patterns for constructing a Zoom archival pipeline.

The Problem #

Zoom is ubiquitous in research: lab meetings, seminars, thesis defenses, collaborator calls, and conference sessions are routinely recorded. But these recordings are fragile:

• Cloud recordings are deleted after a configurable retention period (often 30-120 days) or when storage quotas are exceeded
• Local recordings live on individual laptops and are rarely backed up systematically
• Institutional Zoom accounts may be deprovisioned when personnel leave
• Zoom’s built-in sharing is not designed for long-term preservation

For research groups, losing these recordings means losing institutional knowledge – context, decisions, and discussions that are never written down anywhere else.

Cloud Recordings via the Zoom API #

Zoom provides a REST API that includes endpoints for listing and downloading cloud recordings. This is the most reliable path for automated archival.

API Approach #

# Conceptual workflow -- not a complete implementation
import requests

# 1. Authenticate via Server-to-Server OAuth
access_token = get_zoom_access_token(account_id, client_id, client_secret)

# 2. List recordings for a user or date range
recordings = requests.get(
    "https://api.zoom.us/v2/users/{userId}/recordings",
    headers={"Authorization": f"Bearer {access_token}"},
    params={"from": "2025-01-01", "to": "2025-12-31"}
).json()

# 3. Download each recording file
for meeting in recordings["meetings"]:
    for file in meeting["recording_files"]:
        download_url = file["download_url"] + f"?access_token={access_token}"
        # Download and store in git-annex

Key API Endpoints #

Authentication #

Zoom’s API uses Server-to-Server OAuth apps (recommended) or JWT (deprecated). Setting up a Server-to-Server OAuth app requires:

1. Creating an app in the Zoom App Marketplace
2. Configuring scopes: recording:read:admin (or recording:read for user-level)
3. Obtaining account_id, client_id, and client_secret

Integration with git-annex #

A practical archival script would:

#!/bin/bash
# zoom-archive.sh - Conceptual archival workflow
ARCHIVE_DIR="$1"
cd "$ARCHIVE_DIR" || exit 1

# Run the Python download script
python zoom_download.py --output ./recordings/

# Each meeting gets a directory:
#   recordings/2025-03-15_Lab-Meeting/
#     video.mp4          -> git-annex
#     audio_only.m4a     -> git-annex
#     transcript.vtt     -> git (text, searchable)
#     chat.txt           -> git (text)
#     meeting_info.json  -> git (metadata)

# Add to git-annex
git annex add --include='*.mp4' --include='*.m4a'
git add --all  # transcripts, chat logs, metadata go to git
git commit -m "Zoom archive update $(date -I)"

Local Recording Archival #

For locally saved Zoom recordings (stored on the host machine), the approach is simpler but requires discipline:

Default Local Recording Location #

Zoom saves local recordings to:

• macOS: ~/Documents/Zoom/
• Windows: %USERPROFILE%\Documents\Zoom\
• Linux: ~/Documents/Zoom/

Each recording creates a directory with:

• video.mp4 – the recording video
• audio_only.m4a – audio-only track
• chat.txt – in-meeting chat log (if any)

Scripted Import #

#!/bin/bash
# import-local-zoom.sh - Import local Zoom recordings to git-annex
ZOOM_DIR="$HOME/Documents/Zoom"
ARCHIVE_DIR="$1"

cd "$ARCHIVE_DIR" || exit 1

# Copy new recordings
for meeting_dir in "$ZOOM_DIR"/*/; do
    meeting_name=$(basename "$meeting_dir")
    if [ ! -d "recordings/$meeting_name" ]; then
        cp -r "$meeting_dir" "recordings/$meeting_name"
    fi
done

# Add to git-annex
git annex add recordings/ --include='*.mp4' --include='*.m4a'
git add --all
git commit -m "Import local Zoom recordings $(date -I)"

Zoom’s Built-in Transcripts #

Zoom can generate transcripts for cloud recordings automatically. These are valuable for AI readiness:

• Audio transcript (.vtt format) – time-stamped text of spoken content
• Meeting summary (if AI Companion is enabled) – structured summary of the meeting

When downloading via the API, transcript files are included as separate recording file entries with file_type: "TRANSCRIPT". These should be stored in git (not annex) since they are small text files that benefit from version tracking and full-text search.

Enhancing Transcripts #

Zoom’s auto-generated transcripts can be improved with:

• Whisper (OpenAI) – run on the downloaded audio for higher-quality transcription
• Speaker diarization – identify who said what (Zoom transcripts sometimes include this; Whisper-based tools like whisperx can add it)
• Manual correction – for critical meetings, reviewing and correcting the auto-generated transcript

AI Readiness #

Zoom recordings are ai-manual out of the box, but Zoom’s transcript features significantly improve this:

If Zoom’s transcription is enabled for cloud recordings, a substantial portion of the meeting content becomes text-searchable and AI-accessible immediately upon archival.

Tooling Status #

The maturity level for Zoom archival is concept. There is no single, maintained, open-source tool that provides a complete Zoom-to-git-annex pipeline. The pieces exist:

• Zoom’s API is well-documented and stable
• git-annex handles the storage layer well
• Python libraries for Zoom API interaction exist (though none are specifically designed for archival)

What is missing is a purpose-built tool analogous to annextube for YouTube. This represents an opportunity for the con/serve ecosystem: a con/zoomtube or similar tool that handles authentication, incremental download, transcript extraction, and DataLad integration in a single package.

Existing Partial Solutions #

• zoom-recording-downloader – Python script for bulk downloading cloud recordings via Zoom API
• zoomdl – Go-based Zoom recording downloader (archived)
• Various institutional scripts shared on GitHub – search for “zoom recording backup” or “zoom archive script”

None of these integrate with git-annex or DataLad, but they demonstrate the API patterns needed.

See Also #

• annextube – the model for what a Zoom archival tool could become
• yt-dlp – general video download (does not support Zoom)