plaud-connector

plaud-downloader

CLI tool that bulk-exports transcripts and AI summaries from plaud.ai. Can also trigger transcription for unprocessed recordings.

Plaud doesn't have a public API, so this tool talks to the same endpoints their web app uses (reverse-engineered from browser traffic). It works today, but if they change their API, it might break.

What it does

• Downloads all your recordings as JSON (structured transcript with speaker labels and timestamps) and Markdown (human-readable)
• Organizes files by your Plaud tags/folders
• Triggers transcription for recordings that haven't been processed yet
• Incremental sync: running it again only downloads new stuff

Requirements

• Python 3.10+
• A plaud.ai account with recordings

Installation

git clone https://github.com/Continero/plaud-connector.git
cd plaud-connector
python3 -m venv .venv
source .venv/bin/activate
pip install -e .

After installation, the plaud command is available in your terminal (while the venv is active).

Configuration

Copy the example config and fill in your credentials:

cp .env.example .env

Then edit .env with one of these auth methods:

Option A: Bearer token (Google SSO users, or anyone)

If you signed up with Google, you don't have a password. Grab your token from the browser:

1. Open app.plaud.ai and sign in
2. Open DevTools (Cmd+Option+I on Mac, F12 on Windows)
3. Go to the Network tab
4. Click any recording to trigger a request
5. Find a request to api.plaud.ai, click it, and copy the Authorization header value

PLAUD_TOKEN=bearer eyJhbGciOi...

The token expires after some time. When it does, repeat the steps above to get a fresh one.

Option B: Email and password

If you have a regular account with a password:

[email protected]
PLAUD_PASSWORD=your_password

EU accounts

If your account is on the EU instance (check your browser network tab, the requests go to api-euc1.plaud.ai instead of api.plaud.ai):

PLAUD_API_BASE=https://api-euc1.plaud.ai

Usage

Quick start: sync everything

The sync command does it all in one go. It finds recordings without transcripts, triggers generation, waits for them to finish, and downloads everything.

plaud sync

╔══════════════════════════════════════════════╗
║  plaud-downloader                            ║
║  Bulk export transcripts from plaud.ai       ║
╚══════════════════════════════════════════════╝

Step 1/3  Fetching recordings from plaud.ai...
          165 recordings (188.5 hours total)
          162 with transcripts, 3 without

Step 2/3  Generating missing transcripts...
          Triggering transcription for 2 recordings:
            -> Team standup (3811s)
            -> Project review (1820s)

          Waiting for 2 transcriptions to complete...
            Done: Team standup
            Done: Project review

Step 3/3  Downloading transcripts and summaries...
            + Team standup
            + Project review

          2 new, 163 already up to date

══════════════════════════════════════════════
  Sync complete!
  165 recordings  |  2 exported  |  163 skipped
  Output: /Users/you/plaud-downloader/output/
══════════════════════════════════════════════

Individual commands

If you prefer more control:

plaud list                        # show all recordings
plaud generate --dry-run          # preview what would be transcribed
plaud generate                    # trigger transcription (don't wait)
plaud generate --wait             # trigger and wait for completion
plaud download                    # download to ./output/
plaud download -o ~/plaud-export  # download to custom directory
plaud download --force            # re-download even if files exist

Command options

plaud sync [OPTIONS]
  -o, --output TEXT       Output directory (default: output)
  --min-duration INT      Skip recordings shorter than N seconds (default: 10)
  --wait / --no-wait      Wait for transcription (default: --wait)

plaud generate [OPTIONS]
  --min-duration INT      Skip recordings shorter than N seconds (default: 10)
  --dry-run               Show what would be triggered, don't do it
  --wait / --no-wait      Wait for completion (default: --no-wait)

plaud download [OPTIONS]
  -o, --output TEXT       Output directory (default: output)
  --force                 Re-download even if files already exist

Output format

Each recording produces two files, organized into folders matching your Plaud tags. Untagged recordings go to Unsorted/.

output/
  Work/
    2026-03-17_Team-standup.json
    2026-03-17_Team-standup.md
  Unsorted/
    2026-03-15_Quick-note.json
    2026-03-15_Quick-note.md

JSON

Structured data with transcript segments (speaker, text, timestamps) and the AI-generated summary: