Composable CLI for your meeting workflow
Hack your own way, or pair it with Claude Code Codex OpenClaw .
Brought to you by 
@fastrepl/char
Installation
Available for macOS (x86 and arm).
Want other platforms? PRs welcome at apps/cli .
Or `npx char@latest --help`
Or install SKILL, and let an agent install and set it up for you.
Or use skills.sh without installing char first
char record
Record audio to an MP3 file
▸ Why not just use default voice recorder app?
It can record microphone (--audio input), system audio (--audio output), or both with Echo Cancellation applied (--audio dual).
char transcribe
Transcribe an audio file
Use it for example input
Key needed. Deepgram/Soniox/AssemblyAI are considered stable. Others are in beta.
`char models download whisper-small-en` needed
▸ Looking for realtime transcription, calendar syncing, etc?
We have a desktop app with all of that. Run char desktop.
char models
Manage local models
try with `--format json` too
▸ Is it Whisper.cpp?
Whisper.cpp for all platforms, and additionally Cactus for ARM (e.g. Apple Silicon).
For Agents
char skill installornpx skills add fastrepl/char@char-clito install skill.- Use
--format jsoninstead, if it is not intended for human reading.
Reference
▸Global options
-v, --verbose Increase logging verbosity -q, --quiet Decrease logging verbosity Commands
▸ record Record audio to an MP3 file
char record [--audio] [-o] [-f] [--base] [-h] [-V]
Options
--audio <AUDIO> [dual] -o, --output <FILE> -f, --format <FORMAT> [pretty] --base <DIR> -h, --help Print help -V, --version Print version ▸ transcribe Transcribe an audio file
char transcribe [-i] <-p> [-k] [-o] [-f] [--base] [--base-url] [--api-key] [-m] [-l] [-h] [-V] [target]
Options
-i, --input <FILE> -p, --provider <PROVIDER> -k, --keyword <KEYWORD> -o, --output <FILE> -f, --format <FORMAT> [pretty] --base <DIR> --base-url <BASE_URL> --api-key <API_KEY> -m, --model <MODEL> -l, --language <LANGUAGE> [en] -h, --help Print help -V, --version Print version Arguments
<INPUT> Timestamp (e.g. 20260327_143022) or path to an audio file ▸ models Manage local models
char models [--base] [-h] [-V] <command>
Options
--base <DIR> -h, --help Print help -V, --version Print version Subcommands
▸ list List local models and their status
char models list [-f] [-h] [-V]
Options
-f, --format <FORMAT> [pretty] -h, --help Print help -V, --version Print version ▸ download Download a model by name
char models download [-h] [-V] <name>
Options
-h, --help Print help -V, --version Print version Arguments
<NAME> ▸ delete Delete a downloaded model
char models delete [--force] [-h] [-V] <name>
Options
--force -h, --help Print help -V, --version Print version Arguments
<NAME> ▸ completions Generate shell completions
char completions [-h] [-V] <shell>
Options
-h, --help Print help -V, --version Print version Arguments
<SHELL> ▸ play Play an audio file
char play [--base] [-h] [-V] <target>
Options
--base <DIR> Base directory for session lookup -h, --help Print help -V, --version Print version Arguments
<TARGET> Timestamp (e.g. 20260327_143022) or path to an audio file ▸ skill Install char skill for AI coding agents
char skill [-h] [-V] <command>
Options
-h, --help Print help -V, --version Print version Subcommands
▸ install Install char skill for AI coding agents
char skill install [-f] [-h] [-V]
Options
-f, --format <FORMAT> Skip interactive selection: claude, codex, github-copilot -h, --help Print help -V, --version Print version ▸ update Update char to the latest version
char update [-h] [-V]
Options
-h, --help Print help -V, --version Print version ▸ desktop Open the desktop app or download page
char desktop [-h] [-V]
Options
-h, --help Print help -V, --version Print version