CLI: Audio Commands

Store, transcribe, search, and manage audio memories from the command line.

Updated Oct 6, 2026

Overview

The audio commands let you work with audio memories — recordings, podcasts, meetings, and voice notes. Audio files can be transcribed, searched by transcript content, and analyzed for speaker diarization.

Command aliases: audio, aud

Uploading Audio

Upload audio files using the remember command with the --audio flag. Supported formats include mp3, wav, ogg, flac, m4a, aac, webm, and opus.

# Upload an audio file
trix remember --audio recording.mp3

# Upload with tags and space
trix remember --audio meeting.wav --space work --tags standup,weekly

# Upload with a description
trix remember --audio lecture.m4a "CS 101 intro lecture"

# Upload without auto-transcription
trix remember --audio notes.ogg --transcribe=false

# Upload as private
trix remember --audio personal.mp3 --private

Upload Flags

Flag	Description
`--audio <file>`	Path to the audio file to upload
`--transcribe`	Auto-transcribe after upload (default: true, use `--transcribe=false` to disable)
`-s, --space`	Target space for the audio memory
`-t, --tags`	Comma-separated tags
`--private`	Mark the memory as private
`--priority`	Set memory priority (higher values = more important)

Listing Audio Memories

# List audio memories
trix audio list

# Filter by space
trix audio list --space meetings

# Filter by tags with a custom limit
trix audio list --tags podcast,interview --limit 50

# JSON output
trix audio list -o json

List Flags

Flag	Default	Description
`-s, --space`	default space	Filter by space
`-l, --limit`	20	Maximum number of results
`-t, --tags`	—	Filter by tags (comma-separated)

Audio Info

View detailed metadata about an audio memory, including duration, format, sample rate, channels, bitrate, file size, and transcription status.

trix audio info mem_abc123

Example output:

  ID:            mem_abc123
  Type:          audio
  Duration:      12:34
  Format:        mp3
  Sample Rate:   44100 Hz
  Channels:      stereo
  Bitrate:       192 kbps
  Size:          14.2 MB
  Tags:          meeting, standup

  Transcription: Completed
  Language:      en
  Speakers:      3
  Segments:      87
  View transcript: trix audio transcript mem_abc123

  Created:       2 hours ago
  Updated:       1 hour ago

Searching Audio Transcripts

Search within audio transcripts using natural language queries. Results include similarity scores and which fields matched.

# Search across all audio
trix audio search "project deadline"

# Search within a space
trix audio search "budget discussion" --space meetings

# Limit results
trix audio search "action items" --limit 20

# Filter by tags
trix audio search "quarterly review" --tags finance

Search Flags

Flag	Default	Description
`-s, --space`	default space	Search within a specific space
`-l, --limit`	10	Maximum number of results
`-t, --tags`	—	Filter by tags (comma-separated)

Transcription

Trigger transcription for an audio memory. This starts an asynchronous job that extracts text content from the audio file.

# Trigger transcription
trix audio transcribe mem_abc123

Example output:

✓ Transcription job queued: job_xyz789
  Check status with: trix jobs get job_xyz789

Transcription runs automatically when you upload audio with trix remember --audio (unless you pass --transcribe=false). Use trix audio transcribe to manually trigger it later.

Viewing Transcripts

Retrieve the full transcript for an audio memory. By default, timestamps are included. You can also enable speaker labels.

# Get transcript with timestamps (default)
trix audio transcript mem_abc123

# Plain transcript without timestamps
trix audio transcript mem_abc123 --with-timestamps=false

# Include speaker labels
trix audio transcript mem_abc123 --speaker-labels

Transcript Flags

Flag	Default	Description
`--with-timestamps`	true	Include timestamps in output
`--speaker-labels`	false	Include speaker labels

Speaker Diarization

View detected speakers in an audio memory, including speaking time, segment count, and confidence scores.

trix audio speakers mem_abc123

Example output:

Speakers in mem_abc123

SPEAKER  LABEL  TIME    SEGMENTS  CONFIDENCE
A        -      4:12    23        92%
B        -      3:45    19        88%
C        -      1:30    8         85%

  Total speakers: 3

Playing Audio

Open an audio memory in your system's default audio player, or print the URL for scripting.

# Open in default player
trix audio play mem_abc123

# Just print the URL
trix audio play mem_abc123 --url-only

Play Flags

Flag	Description
`-u, --url-only`	Print the audio URL without opening a player

The play command uses xdg-open on Linux, open on macOS, and rundll32 on Windows to launch the default audio player.

Command Reference

Command	Description
`audio list`	List audio memories with optional filtering
`audio info <id>`	Show audio metadata and transcription status
`audio search <query>`	Search within audio transcripts
`audio transcribe <id>`	Trigger transcription for an audio file
`audio transcript <id>`	Get full transcript with timestamps
`audio speakers <id>`	Show speaker diarization results
`audio play <id>`	Open audio in system default player
`remember --audio <file>`	Upload an audio file as a memory

Next Steps

Video Commands — Manage video memories
Memory Commands — General memory management
Background Jobs — Monitor transcription jobs