Enum Commands 

pub enum Commands {
    Init,
    List,
    Models,
    ModelInfo {
        model: String,
    },
    Chat,
    Version,
    Update {
        check: bool,
        force: bool,
    },
    Status,
    Doctor {
        format: OutputFormat,
    },
    Feedback {
        stdout: bool,
        format: OutputFormat,
    },
    SelfTest {
        format: OutputFormat,
        keep_workspace: bool,
    },
    Tasks {
        limit: usize,
    },
    Task {
        id: String,
        follow: bool,
    },
    Processes {
        limit: usize,
    },
    Logs {
        id: String,
    },
    Stop {
        id: String,
    },
    Restart {
        id: String,
    },
    Open {
        target: String,
    },
    Ports,
    Approvals,
    Approve {
        id: String,
    },
    Deny {
        id: String,
    },
    Cancel {
        id: String,
    },
    ToolRuns {
        limit: usize,
    },
    Checkpoints {
        limit: usize,
    },
    Restore {
        id: String,
        force: bool,
    },
    Plugin {
        command: PluginCommand,
    },
    Daemon {
        command: DaemonCommand,
    },
    Pair {
        command: PairCommand,
    },
    Qa {
        command: QaCommand,
    },
    Add {
        name: String,
        yes: bool,
        command: Option<String>,
        arg: Vec<String>,
        env: Vec<String>,
        url: Option<String>,
        header: Vec<String>,
        env_header: Vec<String>,
    },
    Remove {
        name: String,
    },
    Mcp,
    Pr {
        command: PrCommand,
    },
    CloudSetup,
    Login {
        provider: Option<String>,
    },
    Logout {
        provider: String,
    },
    Run {
        prompt: Option<String>,
        format: OutputFormat,
        max_tokens: Option<usize>,
        no_execute: bool,
        allow_untrusted_tools: bool,
        output_schema: Option<PathBuf>,
        plan: bool,
        plan_autoaccept: bool,
    },
}

Variants

Init

Initialize configuration

List

List available models

Models

List model/provider capability records

ModelInfo

Show static and cached capability info for a model id

Fields

model: String

Model id, e.g. /

Chat

Start a chat session (default)

Version

Show version information

Update

Update Mermaid to the latest release

Fields

check: bool

Only report whether an update is available; don’t install it

force: bool

Reinstall even if already on the latest version

Status

Check status of dependencies and backends

Doctor

Check first-run readiness and explain what Mermaid can do now

Fields

format: OutputFormat

Output format (text, json, markdown)

Feedback

Write a local diagnostic bundle: doctor report, config summary (names and booleans only), recent trace events, and the log tail. Nothing is uploaded — the file stays on this machine.

Fields

stdout: bool

Print to stdout instead of writing mermaid-feedback- in the current directory

format: OutputFormat

Output format (markdown or json)

SelfTest

Run fast deterministic Mermaid self-tests

Fields

format: OutputFormat

Output format (text, json, markdown)

keep_workspace: bool

Keep the temporary self-test workspace after the run

Tasks

List durable runtime tasks

Fields

limit: usize

Maximum number of tasks to show

Task

Show one durable runtime task and its timeline

Fields

id: String

Task id

follow: bool

Attach to the task’s live event stream (daemon required): prints NDJSON RunEvent lines to stdout and exits after the terminal result. Works on queued tasks too (waits for the run to start).

Processes

List Mermaid-managed background processes

Fields

limit: usize

Maximum number of processes to show

Logs

Print a managed process log

Fields

id: String

Process id from mermaid processes

Stop

Stop a managed process

Fields

id: String

Process id from mermaid processes

Restart

Restart a managed process

Fields

id: String

Process id from mermaid processes

Open

Open a URL, file, or managed process URL

Fields

target: String

URL, path, or process id

Ports

Show listening TCP ports

Approvals

List pending approvals

Approve

Approve a pending approval record

Fields

id: String

Approval id

Deny

Deny a pending approval record

Fields

id: String

Approval id

Cancel

Cancel a queued or running daemon task

Fields

id: String

Task id from mermaid tasks

ToolRuns

List recent persisted tool runs

Fields

limit: usize

Maximum number of tool runs to show

Checkpoints

List checkpoints

Fields

limit: usize

Maximum number of checkpoints to show

Restore

Restore a checkpoint by id

Fields

id: String

Checkpoint id

force: bool

Skip the confirmation prompt (required for non-interactive use)

Plugin

Manage Mermaid plugin bundles

Fields

command: PluginCommand

Daemon

Manage Mermaid’s Linux background service

Fields

command: DaemonCommand

Pair

Manage remote pairing tokens

Fields

command: PairCommand

Qa

Internal self-QA commands. Hidden from normal help output.

Fields

command: QaCommand

Add

Add an MCP server (e.g., mermaid add context7)

Fields

name: String

MCP server name (registry key, or a label when using –command)

yes: bool

Skip the confirmation prompt before fetching and running a package that is not in the built-in registry (for scripted/CI use). Without this, adding an unknown package fails closed when there is no TTY.

command: Option<String>

Register a raw command server instead of resolving from the registry: the executable to run (e.g. uvx, node, /path/to/mcp-server).

arg: Vec<String>

Argument for –command, repeatable and order-preserving (e.g. --arg mcp-server-git --arg --repository --arg .).

env: Vec<String>

Environment variable for –command: repeatable KEY=VALUE.

url: Option<String>

Register a remote Streamable HTTP server instead: the MCP endpoint URL (https://..., or http:// to localhost only).

header: Vec<String>

Literal HTTP header for –url: repeatable 'Name: Value' (e.g. --header 'Authorization: Bearer TOKEN').

env_header: Vec<String>

HTTP header for –url whose value is read from an environment variable at request time: repeatable Header=ENV_VAR, so the secret never lands in config.toml.

Remove

Remove a configured MCP server

Fields

name: String

MCP server name to remove

Mcp

List configured MCP servers

Pr

Create a pull/merge request from the current branch via the host CLI (gh for GitHub, glab for GitLab)

Fields

command: PrCommand

CloudSetup

Configure Ollama Cloud API key (interactive prompt). Run this from your shell before starting mermaid — it reads stdin and doesn’t work from inside the TUI.

Login

Store a provider API key in the OS keyring, or list key status

Fields

provider: Option<String>

Provider name (e.g. groq, anthropic, ollama). Omit to list every provider’s key status.

Logout

Remove a provider API key from the OS keyring

Fields

provider: String

Provider name whose stored key to remove

Run

Run a single prompt non-interactively

Fields

prompt: Option<String>

Prompt to execute. Omit or pass - to read it from piped stdin; piped stdin alongside a prompt is appended as a fenced block.

format: OutputFormat

Output format (text, json, markdown, ndjson)

max_tokens: Option<usize>

Maximum tokens to generate

no_execute: bool

Don’t execute agent actions (dry run)

allow_untrusted_tools: bool

Allow non-replayable tools (web/mcp/subagent/computer-use) to run on an Ask decision in this headless run. Off by default — ask mode otherwise refuses them when there’s no approval UI.

output_schema: Option<PathBuf>

JSON Schema file the final answer must conform to. The agentic loop runs normally; one extra formatting turn (no tools, native constrained output where the provider supports it) reshapes the final answer, validated client-side. Failures are reported in the run’s errors; the text answer is still returned.

plan: bool

Run in plan mode: the agent explores read-only and produces a plan file (.mermaid/plans/) instead of making changes. With no approval UI the plan is accepted as the run’s deliverable but implementation does NOT start.

plan_autoaccept: bool

With –plan: the moment the plan is presented, accept it and continue straight into implementation in the same run.

Trait Implementations

impl Debug for Commands

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl FromArgMatches for Commands

fn from_arg_matches(__clap_arg_matches: &ArgMatches) -> Result<Self, Error>

Instantiate Self from ArgMatches, parsing the arguments as needed.

fn from_arg_matches_mut( __clap_arg_matches: &mut ArgMatches, ) -> Result<Self, Error>

Instantiate Self from ArgMatches, parsing the arguments as needed.

fn update_from_arg_matches( &mut self, __clap_arg_matches: &ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

fn update_from_arg_matches_mut<'b>( &mut self, __clap_arg_matches: &mut ArgMatches, ) -> Result<(), Error>

Assign values from ArgMatches to self.

impl Subcommand for Commands

fn augment_subcommands<'b>(__clap_app: Command) -> Command

Append to Command so it can instantiate Self via FromArgMatches::from_arg_matches_mut

fn augment_subcommands_for_update<'b>(__clap_app: Command) -> Command

Append to Command so it can instantiate self via FromArgMatches::update_from_arg_matches_mut

fn has_subcommand(__clap_name: &str) -> bool

Test whether Self can parse a specific subcommand