CLI: Memory Commands

Store, search, boost, and delete memories with the Trix CLI.

Updated Oct 6, 2026

Overview

The CLI provides intuitive commands for working with memories: remember to store, recall to search, reinforce and weaken to adjust priority, and forget to delete. The audio commands provide specialized support for audio memories including transcription, speaker diarization, and search within transcripts. The videos commands enable video memory management with transcription, metadata extraction, thumbnail generation, and transcript search. These commands are designed for both interactive use and scripting.

remember

Store new memories. Aliases: store, save, add

# Store a simple memory
trix remember "User prefers dark mode and compact layouts"

# Store with a specific space
trix remember "Meeting notes from Q4 planning" --space work

# Store with tags
trix remember "Loves hiking in the mountains" --tags outdoor,hobby

# Read from a file
trix remember --file notes.txt

# Read from clipboard
trix remember --clipboard

# Open editor for long content
trix remember --edit

# Read from stdin
echo "Important context" | trix remember -

# Fetch content from URL
trix remember --url https://example.com/doc.html

# Store a bookmark
trix remember --bookmark "https://example.com/article"

# Quick capture modes
trix remember "Quick thought" --note       # Adds #note tag
trix remember "Buy milk" --task            # Adds #task tag
trix remember "const x = 1" --snippet --lang javascript

# Include context
trix remember "Fix for the bug" --git-context  # Include git branch/commit
trix remember "Config notes" --cwd             # Include working directory

# Source tracking
trix remember "Quote from book" --source "Clean Code" --author "Uncle Bob"

# Smart processing
trix remember "Ideas #project #urgent" --extract-tags  # Auto-extract hashtags
trix remember "Important doc" --title "Project Setup"

# Output options
trix remember "data" --id-only    # Just output the ID
trix remember "data" --copy-id    # Copy ID to clipboard
trix remember "data" --quiet      # Suppress output
trix remember "data" --confirm    # Preview before saving
trix remember "data" --open       # Open in browser after creation

# Expiration
trix remember "Temp note" --expire 24h
trix remember "Weekly reminder" --expire 7d

# Privacy
trix remember "Secret" --private

# With template
trix remember --template standup

remember Options

Option	Description
`--space`	Target space for the memory
`--tags`	Comma-separated tags
`--file`	Read content from a file
`--clipboard`	Read content from clipboard
`--edit`	Open editor for content
`--url`	Fetch content from URL
`--bookmark`	Store URL as a bookmark
`--expire`	Set expiration (e.g., 24h, 7d, 30d)
`--private`	Mark memory as private
`--template`	Use a predefined template
`--template-pick, -T`	Interactively select a template from available options
`--session`	Link to a specific session (use "current" for active session)
`--no-session`	Don't auto-link to current session

Auto Session Linking

When you have an active session (set via trix session switch), new memories are automatically linked to that session. This helps organize related work without requiring explicit --session flags on every command.

# Set active session
trix session switch "My Project"

# These memories auto-link to "My Project"
trix remember "Important note"
trix remember "Another note"

# Opt out for a single memory
trix remember "Unrelated thought" --no-session

# Explicit session override
trix remember "Note for other project" --session sess_other123

Auto-session linking only occurs when there's an active current session. Use trix session current to see if a session is active, and --no-session to create memories without session association.

recall

Search and retrieve relevant memories. Alias: find

# Semantic search (hybrid is default - best results)
trix recall "What are the user's preferences?"

# Different search modes
trix recall "dark mode" --mode fulltext   # Keyword matching
trix recall "preferences" --mode semantic  # Vector similarity
trix recall "user settings" --mode hybrid  # Combined (default)

# Output styles
trix recall "query" --style table    # Tabular (default)
trix recall "query" --style full     # Complete content
trix recall "query" --style brief    # Compact ID + first line
trix recall "query" --style context  # For LLM prompts
trix recall "query" --style raw      # Just content

# Filtering
trix recall "notes" --space work
trix recall "hobbies" --tags outdoor
trix recall "query" --min-score 0.7  # Filter by similarity
trix recall "query" --limit 5

# Time filters
trix recall "standup" --today        # Last 24 hours
trix recall "notes" --week           # Last 7 days
trix recall "logs" --since 2h        # Custom duration

# Natural language time filters (NEW)
trix recall "meetings" --since "3 days ago"
trix recall "notes" --since "last week"
trix recall "updates" --since "yesterday" --until "today"
trix recall "reports" --since "this month"
trix recall "2024 planning" --since "last year"

# Sorting
trix recall "query" --sort score     # By relevance (default)
trix recall "query" --sort date      # By creation date
trix recall "query" --sort priority  # By priority

# Get first/best result only
trix recall "api key" --first

# Interactive mode
trix recall "projects" -i            # Browse and select

# Discovery
trix recall --related mem_abc123     # Find similar memories
trix recall --related mem_abc123 --first --copy

# Clipboard
trix recall "api key" --copy         # Copy top result

# Scripting output
trix recall "query" -o json          # JSON output
trix recall "query" --id-only        # Just IDs
trix recall "query" --count          # Just count
trix recall "query" --no-header      # No table header
trix recall "query" --jq '.[].content'  # jq filter

recall Options

Option	Description
`--mode`	Search mode: hybrid (default), fulltext, semantic
`--style`	Output style: table, full, brief, context, raw
`--space`	Filter by space
`--tags`	Filter by tags
`--min-score`	Minimum similarity score (0.0-1.0)
`--limit`	Maximum number of results
`--sort`	Sort by: score, date, priority
`--first`	Return only the first result
`-i`	Interactive selection mode
`--copy`	Copy result to clipboard

Temporal Queries

The --since and --until flags support natural language date expressions:

Option	Description
`today, yesterday`	Relative to current day
`this week, last week`	Week boundaries (Monday-Sunday)
`this month, last month`	Month boundaries
`this year, last year`	Year boundaries
`3 days ago, 2 weeks ago`	Relative durations
`2026-01-15, Jan 15, 2026`	Absolute dates (various formats)

Temporal queries make it easy to find memories from specific time periods without remembering exact dates. Combine with other filters for precise results.

context

Get memories formatted for injection into LLM prompts. Designed for AI agents and automation.

# Get context for a query
trix context "user preferences"

# Limit to token budget
trix context "project info" --max-tokens 2000

# Different output formats
trix context "api docs" --format plain      # Raw content (default)
trix context "api docs" --format xml        # XML-tagged
trix context "api docs" --format numbered   # Numbered list with IDs
trix context "api docs" --format markdown   # Markdown headers

# Filter results
trix context "notes" --space work
trix context "notes" --limit 5
trix context "notes" --min-score 0.5

# Use in a script/agent
CONTEXT=$(trix context "user preferences" --max-tokens 1000)
echo "Based on this context: $CONTEXT"

The context command returns silently (no spinner) and outputs empty for no results, making it ideal for scripting and agent workflows.

reinforce

Boost the priority of important memories. Aliases: boost, strengthen

# Reinforce a single memory
trix reinforce mem_abc123

# Reinforce multiple memories
trix reinforce mem_abc123 mem_def456 mem_ghi789

# Control boost amount
trix reinforce mem_abc123 --boost 3    # Boost by 3 (default is 1)
trix reinforce mem_abc123 --priority 8  # Set explicit priority

# Add tags while reinforcing
trix reinforce mem_abc123 --tags important

# From stdin (batch mode)
trix recall "project notes" --id-only | trix reinforce --batch

# Query-based reinforcement
trix reinforce --query "important meeting" --limit 5

# Interactive mode
trix reinforce -i --query "project notes"

# Preview changes
trix reinforce --query "notes" --dry-run

# Include a reason (stored in audit trail)
trix reinforce mem_abc123 --reason "Confirmed accurate by user"

# JSON output for scripting
trix reinforce mem_abc123 --json

Higher priority memories appear more prominently in search results. Use reinforce to boost memories that are frequently useful or important.

weaken

Reduce the priority of memories. Aliases: decay, dim

# Weaken a memory
trix weaken mem_abc123

# Control decay amount
trix weaken mem_abc123 --decay 2       # Reduce by 2 (default is 1)
trix weaken mem_abc123 --priority 2    # Set explicit priority

# From stdin
trix recall "old notes" --id-only | trix weaken --batch

# Query-based weakening
trix weaken --query "deprecated api" --limit 5
trix weaken --query "2023 meeting" --dry-run

# Interactive mode
trix weaken -i --query "outdated"

forget

Delete memories. Aliases: delete, remove, rm

# Delete a single memory
trix forget mem_abc123

# Delete multiple memories
trix forget mem_abc123 mem_def456

# Skip confirmation
trix forget mem_abc123 --force

# Preview what would be deleted
trix forget mem_abc123 --dry-run

# Delete all expired memories
trix forget --expired --force

# Delete old memories
trix forget --older-than 90d --dry-run
trix forget --older-than 30d --space archive --force

# Delete by tags
trix forget --tags temp,draft --force

# Pipe from recall
trix recall "old notes" --id-only | trix forget -

The forget command permanently deletes memories. Use --dry-run to preview what will be deleted before running with --force.

Memory Management

Detailed memory operations using the memories command. Aliases: mem, m

# List all memories
trix memories list

# List with filters
trix memories list --space work --limit 20
trix memories list --tags important

# Get a specific memory
trix memories get mem_abc123

# Get with quality score
trix memories get mem_abc123 --show-quality

# Create a memory (use remember)
trix remember "New memory content"

# Update a memory
trix memories update mem_abc123 --content "Updated content"
trix memories update mem_abc123 --tags newtag

# Delete a memory
trix forget mem_abc123

# Search memories
trix memories search "query" --mode semantic

# List by minimum quality score
trix memories list --min-quality=0.7

# List pinned memories
trix memories list --pinned

# List protected memories
trix memories list --protected

Batch Operations

# Add tags to memories matching a query
trix batch tag --query "meeting notes" --add important

# Remove tags from memories
trix batch tag --query "old" --remove temp,draft

# Add and remove tags together
trix batch tag --query "draft posts" --remove draft --add published

# Pipe IDs from recall
trix recall "project" --id-only | trix batch tag - --add archived

# Preview changes
trix batch tag --query "draft" --add reviewed --dry-run

Audio Commands

Work with audio memories including playback, transcription, and speaker analysis. Audio memories can be created with the remember command and managed with dedicated audio subcommands.

Uploading Audio

Use the remember command with audio-specific flags to upload audio files.

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

# Upload with a specific space
trix remember --audio meeting.wav --space work

# Upload with tags
trix remember --audio interview.m4a --tags interview,research

# Upload and automatically trigger transcription
trix remember --audio podcast.mp3 --transcribe

# Upload with transcription and additional options
trix remember --audio lecture.wav --transcribe --tags education,notes

Audio Upload Options

Option	Description
`--audio <file>`	Upload an audio file (supports mp3, wav, m4a, ogg, flac)
`--transcribe`	Automatically trigger transcription after upload
`--space`	Target space for the audio memory
`--tags`	Comma-separated tags

Listing Audio Memories

# List all audio memories
trix audio list

# List audio in a specific space
trix audio list --space work

# Limit results
trix audio list --limit 10

# Filter by tags
trix audio list --tags meeting,important

# Output in different formats
trix audio list -o json
trix audio list -o table
trix audio list -o yaml

Audio Information

Get detailed metadata and transcription status for an audio memory.

# Show audio metadata and transcription status
trix audio info aud_abc123

# Output as JSON
trix audio info aud_abc123 -o json

# Show info for multiple audio files
trix audio info aud_abc123 aud_def456

Transcription

Transcribe audio files and retrieve transcripts with timestamps and speaker labels.

# Trigger transcription for an audio file
trix audio transcribe aud_abc123

# Get the full transcript
trix audio transcript aud_abc123

# Get transcript with timestamps
trix audio transcript aud_abc123 --with-timestamps

# Get transcript with speaker labels
trix audio transcript aud_abc123 --speaker-labels

# Get transcript with both timestamps and speaker labels
trix audio transcript aud_abc123 --with-timestamps --speaker-labels

# Output transcript as JSON
trix audio transcript aud_abc123 -o json

# Output transcript as YAML
trix audio transcript aud_abc123 --format yaml

Searching Audio

Search within audio transcripts to find specific content.

# Search within audio transcripts
trix audio search "quarterly revenue"

# Search in a specific space
trix audio search "action items" --space work

# Limit search results
trix audio search "project update" --limit 5

# Filter by tags
trix audio search "deadline" --tags meeting

# Output search results as JSON
trix audio search "budget" -o json

Speaker Diarization

View speaker information and diarization data for transcribed audio.

# Show speaker diarization
trix audio speakers aud_abc123

# Output as JSON
trix audio speakers aud_abc123 -o json

# Output as YAML
trix audio speakers aud_abc123 --format yaml

Playback

Open audio files in your system's default audio player.

# Open audio in system player
trix audio play aud_abc123

# Get only the URL (useful for scripting)
trix audio play aud_abc123 --url-only

Audio Command Options

Option	Applies To	Description
`--space`	list, search	Filter by space
`--limit`	list, search	Limit number of results
`--tags`	list, search	Filter by tags
`--format`	list, info, transcript, search, speakers	Output format: json, table, yaml
`--with-timestamps`	transcript	Include timestamps in transcript output
`--speaker-labels`	transcript	Include speaker labels in transcript output
`--url-only`	play	Only output the audio URL (for scripting)

Transcription is processed asynchronously. Use trix audio info to check the transcription status, or use --transcribe with the upload command to automatically queue transcription.

Speaker diarization automatically identifies different speakers in your audio. This is especially useful for meeting recordings, interviews, and podcasts. Use --speaker-labels with the transcript command to see who said what.

Memory Pinning

Pin memories to prevent them from decaying over time. Pinned memories maintain their priority and are always included in relevant search results.

# Pin a memory
trix memories pin mem_abc123

# Unpin a memory
trix memories unpin mem_abc123

# Set protection level
trix memories update mem_abc123 --protection soft   # Warns before deletion
trix memories update mem_abc123 --protection hard   # Prevents deletion entirely
trix memories update mem_abc123 --protection none   # Remove protection

# List pinned memories
trix memories list --pinned

# List protected memories
trix memories list --protected

# Combine filters
trix memories list --pinned --space work
trix memories list --protected --tags important

Pinning Options

Option	Description
`--level`	Protection level: none, soft, hard
`--pinned`	Filter to show only pinned memories
`--protected`	Filter to show only protected memories

Pinned memories are ideal for core preferences, important facts, or frequently-accessed information that should always be readily available.

Soft Delete

Soft delete allows you to remove memories temporarily while retaining the ability to restore them. This is useful for cleaning up without permanently losing information.

# Soft delete a memory
trix memories delete mem_abc123 --soft

# Restore a soft-deleted memory
trix memories restore mem_abc123

# Permanently delete (standard behavior)
trix forget mem_abc123 --force

# Restore multiple memories
trix memories restore mem_abc123 mem_def456 mem_ghi789

Soft-deleted memories are automatically purged after 30 days. To permanently retain a memory, restore it before the purge period ends.

Topic Extraction

Trix uses LLM-powered topic extraction to automatically categorize and tag memories. Topics help with discovery and organization of related content.

# Get topics for a memory
trix memories topics mem_abc123

# Refresh topics (force re-extraction)
trix memories topics mem_abc123 --refresh

# Filter by category
trix memories topics mem_abc123 --category=TECHNOLOGY
trix memories topics mem_abc123 --category=TECHNOLOGY,BUSINESS

# Filter by minimum relevance
trix memories topics mem_abc123 --min-relevance=0.5

# Combine filters
trix memories topics mem_abc123 --category=TECHNOLOGY --min-relevance=0.7

# Output as JSON for scripting
trix memories topics mem_abc123 -o json

Topic Categories

Category	Description
`TECHNOLOGY`	Programming, software, hardware, technical concepts
`BUSINESS`	Work, projects, meetings, professional topics
`PERSONAL`	Preferences, hobbies, personal information
`REFERENCE`	Documentation, facts, reference material
`TEMPORAL`	Dates, deadlines, scheduled events

Topic Options

Option	Description
`--refresh`	Force re-extraction of topics
`--category`	Filter by topic category (comma-separated)
`--min-relevance`	Minimum relevance score (0.0-1.0)

Quality and Enrichment

Trix automatically calculates quality scores for memories and can enrich them with LLM-generated metadata including summaries, entities, and topics.

# Get quality score for a memory
trix memories get mem_abc123 --show-quality

# List memories by minimum quality score
trix memories list --min-quality=0.7

# Sort by quality
trix memories list --sort=quality

# Trigger enrichment for a memory
trix memories enrich mem_abc123

# Enrich with specific operations
trix memories enrich mem_abc123 --operations=topics,summary
trix memories enrich mem_abc123 --operations=entities,quality

# Batch enrich memories
trix recall "project notes" --id-only | xargs -I {} trix memories enrich {}

# Force re-enrichment
trix memories enrich mem_abc123 --force

Enrichment Operations

Operation	Description
`topics`	Extract and categorize topics from content
`summary`	Generate a concise summary of the memory
`entities`	Extract named entities (people, places, organizations)
`quality`	Calculate a quality score based on content analysis

Quality scores range from 0.0 to 1.0, with higher scores indicating more complete, well-structured, and useful content. Use --min-quality to filter out lower-quality memories from search results.

Cluster Commands

Clusters automatically group related memories at different scales, helping you discover patterns and navigate large memory collections.

# List clusters at different scales
trix cluster list                        # Default scale
trix cluster list --scale=fine           # Small, specific clusters
trix cluster list --scale=medium         # Balanced clusters
trix cluster list --scale=coarse         # Large, broad clusters

# Get cluster details
trix cluster get clus_abc123

# Get cluster with member memories
trix cluster get clus_abc123 --include-memories

# Search with cluster expansion
trix recall "project ideas" --expand
trix recall "quarterly goals" --expand

# List clusters with limits
trix cluster list --scale=fine --limit=10

# Output as JSON
trix cluster list -o json

# Trigger cluster generation
trix cluster generate                        # Generate for all memories
trix cluster generate --space work           # Generate for specific space
trix cluster generate --ids mem_1,mem_2,mem_3  # Generate for specific memories

Cluster Scales

Scale	Description
`fine`	Small, tightly-related clusters (3-10 memories)
`medium`	Balanced clusters with moderate similarity (10-50 memories)
`coarse`	Large, thematically-related clusters (50+ memories)

Cluster Options

Option	Description
`--scale`	Cluster scale: fine, medium, coarse
`--include-memories`	Include member memories in cluster details
`--cluster-scale`	Filter search results by cluster scale
`--limit`	Maximum number of clusters to return

Clusters are automatically updated as you add, modify, or delete memories. Use different scales depending on whether you need precise groupings or broad themes.

Relationship Graph

Visualize memory relationships as a graph. Aliases: rel graph

# Show ASCII tree graph (default)
trix relationships graph mem_abc123

# Generate GraphViz DOT format
trix relationships graph mem_abc123 --format dot

# Generate Mermaid diagram
trix relationships graph mem_abc123 --format mermaid

# Control traversal depth
trix relationships graph mem_abc123 --depth 3    # Default is 2

# Filter by relationship types
trix relationships graph mem_abc123 --types supports,related_to

# Include labels in output
trix relationships graph mem_abc123 --labels

# Save DOT output to file for rendering
trix relationships graph mem_abc123 --format dot > graph.dot
dot -Tpng graph.dot -o graph.png

Graph Formats

Format	Description
`ascii`	Human-readable tree visualization in the terminal (default)
`dot`	GraphViz DOT format for rendering with graphviz tools
`mermaid`	Mermaid diagram format for documentation and web rendering

Graph Options

Option	Description
`--format`	Output format: ascii, dot, mermaid
`--depth`	Maximum traversal depth (default: 2)
`--types`	Comma-separated relationship types to include
`--labels`	Include memory content labels in output

The graph visualization helps understand how memories are connected. Use DOT format with GraphViz or Mermaid format for documentation to create visual diagrams.

Typed Relationships

Create semantic relationships between memories with specific types. Aliases: relate, rel

# Create a typed relationship
trix relate mem_abc123 mem_def456 --type depends_on

# Available relationship types
trix relate mem_a mem_b --type depends_on    # A depends on B
trix relate mem_a mem_b --type implements    # A implements B
trix relate mem_a mem_b --type references    # A references B
trix relate mem_a mem_b --type supersedes    # A supersedes (replaces) B
trix relate mem_a mem_b --type related_to    # A is related to B
trix relate mem_a mem_b --type derived_from  # A is derived from B

# Add a note to the relationship
trix relate mem_abc mem_def --type implements \
  --note "Implements the authentication flow from ADR-001"

# List relationships for a memory
trix relationships list mem_abc123
trix relationships list mem_abc123 --direction outgoing
trix relationships list mem_abc123 --direction incoming
trix relationships list mem_abc123 --type depends_on

Relationship Types

Type	Meaning	Example
`depends_on`	Source requires target	Service depends on database
`implements`	Source implements target	Code implements design decision
`references`	Source mentions target	Meeting notes reference project
`supersedes`	Source replaces target	New decision supersedes old one
`related_to`	General association	Topics in the same domain
`derived_from`	Source is based on target	Summary derived from meeting notes

Typed relationships build a semantic knowledge graph. Use trix graph export to visualize these connections, or search with --types to filter by relationship type.

Templates

Create reusable templates for memory creation. Aliases: template, tpl

Built-in templates include:

standup - Daily standup format
meeting - Meeting notes template
idea - Idea capture template
bug - Bug report template
todo - Todo list template

# List all templates
trix templates list

# Show template content
trix templates show standup

# Create a new template
trix templates create weekly

# Edit a template
trix templates edit standup

# Delete a template
trix templates delete mytemplate

# Use a template with remember
trix remember --template standup
trix remember --template meeting --space work

# Interactively pick a template
trix remember --template-pick
trix remember -T                  # Short form

Video Commands

Store, search, and manage video memories with full transcription support. Video commands enable you to upload videos, extract transcripts, search within video content, and manage video metadata.

Uploading Videos

Use the remember command with the --video flag to upload video files.

# Upload a video file
trix remember --video recording.mp4

# Upload and automatically transcribe
trix remember --video meeting.mp4 --transcribe

# Upload with transcription in specific language
trix remember --video presentation.webm --transcribe --language es

# Upload, transcribe, and wait for completion
trix remember --video interview.mov --transcribe --wait

# Upload to a specific space with tags
trix remember --video demo.mp4 --space work --tags presentation,demo

# Upload with metadata
trix remember --video lecture.mp4 --title "CS101 Lecture 3" --tags education

Video Upload Options

Option	Description
`--video`	Path to the video file to upload
`--transcribe`	Automatically trigger transcription after upload
`--language`	Transcription language (ISO code, e.g., en, es, fr, de)
`--wait`	Wait for transcription to complete before returning

Listing Videos

# List all video memories
trix videos list

# Filter by space
trix videos list --space work

# Limit results
trix videos list --limit 10

# Filter by tags
trix videos list --tags meeting,presentation

# Output in different formats
trix videos list -o json
trix videos list --format table
trix videos list --format yaml

Video Information

# Show video metadata (resolution, fps, codec, duration)
trix videos info vid_abc123

# Output as JSON
trix videos info vid_abc123 -o json

# Show detailed metadata
trix videos metadata vid_abc123

# Get detailed technical information
trix videos metadata vid_abc123 --format yaml

Video Transcription

Transcription converts the audio track of videos into searchable text, enabling semantic search within video content.

# Trigger transcription for a video
trix videos transcribe vid_abc123

# Transcribe with specific language
trix videos transcribe vid_abc123 --language fr

# Wait for transcription to complete
trix videos transcribe vid_abc123 --wait

# Get the transcript
trix videos transcript vid_abc123

# Get transcript with timestamps
trix videos transcript vid_abc123 --timestamps

# Save transcript to file
trix videos transcript vid_abc123 --output transcript.txt

# Get transcript in different formats
trix videos transcript vid_abc123 -o json
trix videos transcript vid_abc123 --format yaml

Searching Videos

Search within video transcripts to find specific content across your video memories.

# Search within video transcripts
trix videos search "quarterly revenue"

# Search with filters
trix videos search "project update" --space work
trix videos search "demo" --tags presentation
trix videos search "meeting notes" --limit 5

# Search with timestamps in results
trix videos search "action items" --timestamps

# Output in different formats
trix videos search "budget" -o json
trix videos search "roadmap" --format table

Video Thumbnails

# Get video thumbnail
trix videos thumbnail vid_abc123

# Get thumbnail URL only
trix videos thumbnail vid_abc123 --url-only

# Save thumbnail to file
trix videos thumbnail vid_abc123 --output thumb.jpg

# Get thumbnail at specific timestamp
trix videos thumbnail vid_abc123 --timestamp 30s
trix videos thumbnail vid_abc123 --timestamp 1m30s
trix videos thumbnail vid_abc123 --timestamp 01:30:00

Playing Videos

# Open video in system player
trix videos play vid_abc123

# Get video URL only (for scripting)
trix videos play vid_abc123 --url-only

Video Commands Reference

Command	Description
`trix videos list`	List video memories
`trix videos info <id>`	Show video metadata (resolution, fps, codec, duration)
`trix videos search <query>`	Search within video transcripts
`trix videos transcribe <id>`	Trigger video transcription
`trix videos transcript <id>`	Get video transcript
`trix videos thumbnail <id>`	Get or save video thumbnail
`trix videos metadata <id>`	Show detailed video metadata
`trix videos play <id>`	Open video in system player

Video Flags Reference

Option	Applies To	Description
`--space`	list, search	Filter by space
`--limit`	list, search	Limit number of results
`--tags`	list, search	Filter by tags (comma-separated)
`--format`	list, info, search, transcript, metadata	Output format (json, table, yaml)
`--timestamps`	transcript, search	Include timestamps in output
`--output`	thumbnail, transcript	Save output to file
`--url-only`	thumbnail, play	Only show URL (no download/open)
`--timestamp`	thumbnail	Extract thumbnail at specific time
`--language`	transcribe, remember --video	Transcription language (ISO code)
`--wait`	transcribe, remember --video --transcribe	Wait for transcription to complete

Video Workflow Examples

# Complete workflow: Upload, transcribe, and search
trix remember --video meeting-2026-01-15.mp4 --space work --tags meeting \
  --transcribe --wait
trix videos search "action items" --space work

# Batch transcribe multiple videos
for vid in $(trix videos list -o json | jq -r '.[].id'); do
  trix videos transcribe "$vid" --language en
done

# Export transcript for documentation
trix videos transcript vid_abc123 --timestamps --output docs/meeting-transcript.md

# Create thumbnail gallery
trix videos list -o json | jq -r '.[].id' | while read id; do
  trix videos thumbnail "$id" --output "thumbnails/$id.jpg"
done

# Find videos mentioning specific topics
trix videos search "Q4 planning" -o json | jq '.[] | {id, timestamp, text}'

Video transcription is performed asynchronously. Use --wait to block until transcription completes, or check status later with trix videos info. Transcription quality depends on audio clarity and language support.

For best search results, always transcribe your videos. Transcripts enable semantic search, allowing you to find videos by what was said, not just by title or tags.

Next Steps

Spaces - Organize memories into isolated spaces
Advanced - Export, import, and scripting patterns
Memory Concepts - Understanding memory types and features