plaud

name: plaud description: Use when asked to upload to Plaud, transcribe audio, download recordings, list tags, check analysis status, manage speakers, or work with Plaud API for meeting transcription.

Plaud API Integration

Skill for working with Plaud AI meeting recordings using the plaud-api Python package.

Prerequisites

Install the package:

pip install plaud-api

Authenticate (one-time):

plaud auth setup    # Paste token from browser DevTools

Default Behavior

CRITICAL: Don't ask — run the appropriate command immediately.

Core Commands

List recordings

plaud recordings list
plaud recordings list -n 50

Upload and analyze

plaud recordings upload audio.mp3 --name "Meeting Name" --analyze --language en

Get transcription

plaud transcription get <file_id>
plaud transcription get <file_id> --json

Get AI summary

plaud transcription summary <file_id>

Start analysis manually

plaud transcription start <file_id> --language en
plaud transcription status <file_id>

Download audio

plaud recordings download <file_id>
plaud recordings download <file_id> --url-only

List tags

plaud tags list
plaud tags recordings <tag_id>

Speaker management

plaud speakers list                                            # list known speakers
plaud speakers recording <file_id>                             # who spoke in a recording
plaud speakers rename <file_id> "Speaker 1" "John Smith"       # rename in a recording

Python API (for advanced workflows)

from plaud import PlaudClient

client = PlaudClient()

# List unanalyzed recordings
recordings = client.recordings.list(limit=50)
unanalyzed = [r for r in recordings if not r.has_transcription]

# Batch analyze
for rec in unanalyzed:
    client.transcriptions.start(rec.id, language="en")
    result = client.transcriptions.wait(rec.id, timeout=600)
    client.transcriptions.save_results(rec.id, result)

Troubleshooting

• "Plaud token not found" — run plaud auth setup or set PLAUD_TOKEN env var
• 401 Unauthorized — token expired, run plaud auth setup again
• Analysis timeout — increase timeout: client.transcriptions.wait(id, timeout=900)