meeting-ink-cli 0.1.0

Agent-friendly CLI for Meeting Ink — transcription, summaries, and translations

meeting-ink-cli

Agent-friendly CLI for Meeting Ink — transcription, summaries, and translations.

Designed for both human users and AI agents (Claude Code, Codex, etc.).

Installation

uv (recommended)

uv tool install meeting-ink-cli
uv tool install 'meeting-ink-cli[audio]'      # audio duration detection
uv tool install 'meeting-ink-cli[mcp]'        # MCP server mode
uv tool install 'meeting-ink-cli[audio,mcp]'  # everything

Homebrew

brew tap DeepWaveLab/mi_cli https://github.com/DeepWaveLab/mi_cli
brew install meeting-ink-cli

pip

pip install meeting-ink-cli
pip install 'meeting-ink-cli[audio]'      # audio duration detection
pip install 'meeting-ink-cli[mcp]'        # MCP server mode

Update

# uv
uv tool upgrade meeting-ink-cli

# Homebrew
brew update && brew upgrade meeting-ink-cli

# pip
pip install --upgrade meeting-ink-cli

Quick Start

# Authenticate (auto-extract from browser cookies)
meeting-ink auth

# Or paste token directly / from stdin
meeting-ink auth eyJhbG...
echo $TOKEN | meeting-ink auth -

# List meetings
meeting-ink list

# Show meeting details
meeting-ink show 123

# View transcript
meeting-ink transcript 123

# Generate summary
meeting-ink summary 123 --generate --language zh-TW

# Translate transcript
meeting-ink translate 123 -l en

# Download artifacts
meeting-ink download 123 -t summary -o summary.md

# Upload audio for transcription
meeting-ink upload recording.mp3

# Import from YouTube
meeting-ink from-url "https://youtube.com/watch?v=..."

# Check quota
meeting-ink quota

Commands

Command	Description
auth [token]	Save API token (stdin -, browser auto-extract, or interactive prompt)
whoami	Show user info and subscription
quota	Show transcription quota and usage
list	List meetings (--page, --all, --limit)
show <id>	Meeting detail with transcripts, summaries, translations
upload <file>	Upload audio file (--duration, --name, --wait)
from-url <url>	Import from YouTube/podcast URL
transcript <id>	View transcript (--output to save to file)
summary <id>	View/generate summary (--topic, --template, --language, --generate)
translate <id>	View/generate translation (-l language required)
download <id>	Download artifacts (-t type, --output, --language, --force)
rename <id> <name>	Rename a meeting
delete <id>	Delete a meeting (--force for non-interactive)
serve	Start MCP server on stdio

Transcripts, Summaries & Translations

# View transcript (checks meeting status first)
meeting-ink transcript 123
meeting-ink transcript 123 -o transcript.json

# View existing summary (pinned first, then most recent)
meeting-ink summary 123

# Generate new summary with template and language
meeting-ink summary 123 --generate --template standup --language en
meeting-ink summary 123 --topic "Action items" --per-speaker

# Translate transcript to another language
meeting-ink translate 123 -l en
meeting-ink translate 123 -l ja --no-wait

# Download any artifact
meeting-ink download 123 -t recording -o meeting.mp3
meeting-ink download 123 -t transcript -o transcript.json
meeting-ink download 123 -t summary -o summary.md
meeting-ink download 123 -t translation -l en -o translation.json

Summary Templates

general_meeting (default), standup, one_on_one, project_team_meeting, investor_update, expert_consultation, customer_discovery, customer_qbr, business_progress, bant, project_kickoff, user_research, ideate, issue_resolution

Supported Languages

en-US, zh-TW, ja-JP, de-DE, fr-FR, ru-RU, es-ES, it-IT, th-TH

Shortcuts: en, zh, ja, de, fr, ru, es, it, th

Upload

Supported formats: mp3, wav, m4a, aac, ogg, flac, wma, webm.

Files over 50 MB are automatically chunked. Maximum file size: 2 GB.

# With automatic duration detection (requires mutagen)
meeting-ink upload recording.mp3

# Manual duration (seconds) if mutagen not installed
meeting-ink upload recording.mp3 --duration 3600

# Custom name, don't wait for transcription
meeting-ink upload recording.mp3 --name "Team standup" --no-wait

Agent Usage

All commands support --json for machine-readable output:

meeting-ink list --json
# {"ok": true, "data": {"total": 5, "data": [...]}}

meeting-ink transcript 123 --json
# {"ok": true, "data": {"meeting_id": 123, "content": [...]}}

meeting-ink summary 123 --generate --json
# {"ok": true, "data": {"meeting_id": 123, "summary_id": 456, "content": "..."}}

Use --help-json for structured command metadata:

meeting-ink --help-json         # all commands
meeting-ink summary --help-json # command params

Environment Variables

Variable	Description
MEETING_INK_TOKEN	Auth token (takes precedence over config file)
MEETING_INK_OUTPUT	Force output mode: json or human

Output Format

{"ok": true, "data": ...}
{"ok": false, "error": {"code": "auth_error", "message": "..."}}

Exit Codes

Code	Meaning
0	Success
1	API error (4xx/5xx)
2	Validation error (bad input, missing token)
3	Network error (timeout, unreachable)

MCP Server

Run as a Model Context Protocol server for AI agent integration:

uv pip install 'meeting-ink-cli[mcp]'
meeting-ink serve

Configuration

Token is stored at ~/.config/meeting-ink/config.json with 0600 permissions.

The MEETING_INK_TOKEN environment variable takes precedence over the config file.

Development

git clone <repo>
cd mi_cli
uv sync

# Run tests
uv run pytest -v

# Run linting
uv run ruff check mi_cli/ tests/
uv run ruff format --check mi_cli/ tests/

# Run integration tests (requires MEETING_INK_TOKEN)
MEETING_INK_TOKEN=... uv run pytest -m integration

Architecture

mi_cli/
  cli.py        # Click commands (14 commands)
  client.py     # httpx API client (cookie auth, retry, rate limit)
  config.py     # Token storage, browser cookie extraction
  output.py     # JSON envelope, TTY detection, Rich tables
  validate.py   # Input sanitization, language validation
  upload.py     # Azure Blob upload (single + chunked)
  polling.py    # Adaptive polling for async operations
  help_json.py  # --help-json for agent discoverability
  serve.py      # MCP server (optional)

License

MIT