CLI Flags

Usage

yaku [flags] [text]

Text can be provided as positional arguments, via -f (files), or piped through stdin. See Input & Output for details.

Translation flags

Flag	Type	Default	Description
`--to`	`string`	config `default-target`	Required. Target language code (e.g., `zh-TW`, `ja`, `en`).
`--from`	`string`	auto-detect	Source language code. Omit to let the LLM detect automatically.
`--mode`	`string`	`translate`	`translate` — translate to target language. `polish` — refine writing already in the target language. See Polish Mode.
`--context`	`string`		Domain hint for the LLM (e.g., `"Kubernetes docs"`, `"formal email"`). Helps resolve ambiguous terms.

Input / output flags

Flag	Short	Type	Default	Description
`--file`	`-f`	`string[]`		Input file path(s). Repeat for multiple files: `-f a.md -f b.md`.
`--output`	`-o`	`string`	stdout	Write translated text to a file instead of stdout.
`--format`		`string`	auto-detect	Force input format: `text`, `md`, `json`, `yaml`. Auto-detected from the `-f` file extension when omitted.
`--stream`		`bool`	`false`	Stream output to stdout as the LLM generates it. Text and Markdown formats only.

Backend flags

Flag	Type	Default	Description
`--backend`	`string`	`hosted`	LLM backend: `hosted`, `gemini`, `openai`, `anthropic`. See Backends.
`--model`	`string`	`backend` default	Override the model name (e.g., `gemini-2.5-flash`, `gpt-4o`, `claude-sonnet-4-5-20250514`).
`--api-base`	`string`		API base URL for OpenAI-compatible providers (e.g., `https://api.groq.com/openai/v1`).

Glossary flags

Flag	Type	Default	Description
`--glossary`	`string`	auto-detect	Path to a glossary YAML file. Overrides auto-detection. See Glossary.
`--no-glossary`	`bool`	`false`	Skip glossary loading entirely, even if a glossary file exists in the default locations.

Prompt flags

Flag	Type	Default	Description
`--prompt`	`string`	built-in	Path to a custom system prompt file. Replaces the entire built-in prompt. See Custom Prompts.

Diagnostic flags

Flag	Type	Default	Description
`--verbose`	`bool`	`false`	Print model name, token counts, and elapsed time to stderr after translation.

Example --verbose output (on stderr):

Model: gemini-2.5-flash | Tokens: 42 in / 38 out | Time: 1.2s

Subcommands

`yaku config`

Manage the configuration file at ~/.config/yaku/config.yaml.

Command	Description
`yaku config set <key> <value>`	Set a configuration value.
`yaku config get <key>`	Read a configuration value (file only, no env overlay).
`yaku config path`	Print the config file path.

See Configuration for all available keys.

`yaku login`

Authenticate with the hosted service using a browser-based device authorization flow. Opens your browser with a pre-filled authorization code — just sign in and click Approve. Credentials are stored in ~/.config/yaku/.credentials.

yaku login
# Opens your browser to complete authentication

`yaku logout`

Revoke your session on the server and delete local credentials. If the server is unreachable, credentials are kept so you can retry.

yaku logout

# Force-delete local credentials even if server revocation fails
yaku logout --force

Flag	Description
`--force`	Delete local credentials regardless of whether the server confirmed revocation. The session may remain valid on the server until it expires.

`yaku whoami`

Show the current authentication status by querying the hosted service.

yaku whoami
# Logged in as [email protected] (github)

If not logged in:

Not logged in (using anonymous quota)

If the session is expired or invalid:

Session expired or invalid. Run yaku login to re-authenticate.

If the server is unreachable:

hosted service unreachable. Check your network connection

`yaku quota`

Show your current usage quota for the hosted service. Displays your plan tier, daily and monthly request counts, and the per-request character limit. When using a local backend, reports that no quota applies.

See Hosted Service & Plans for plan tiers and limits.

`yaku version`

Print the CLI version.

yaku version
# yaku v0.1.0