Struct Client 

pub struct Client { /* private fields */ }

An MCP client instance.

Clients are built using ClientBuilder and can connect to servers via various transports (stdio subprocess, SSE, WebSocket).

Implementations

impl Client

pub fn stdio(command: &str, args: &[&str]) -> Result<Client, McpError>

Creates a client connecting to a subprocess via stdio.

Arguments

• command - The command to run (e.g., “uvx”, “npx”)
• args - Arguments to pass to the command

Errors

Returns an error if the subprocess fails to start or initialization fails.

pub fn stdio_with_cx( command: &str, args: &[&str], cx: Cx, ) -> Result<Client, McpError>

Creates a client with a provided Cx for cancellation support.

pub fn builder() -> ClientBuilder

Creates a new client builder.

pub fn ensure_initialized(&mut self) -> Result<(), McpError>

Ensures the client is initialized.

In auto-initialize mode, this performs the initialization handshake on first call. In normal mode, this is a no-op since the client is already initialized.

Since this method takes &mut self, Rust’s borrowing rules guarantee exclusive access, so no additional synchronization is needed.

Errors

Returns an error if initialization fails.

pub fn is_initialized(&self) -> bool

Returns whether the client has been initialized.

pub fn server_info(&self) -> &ServerInfo

Returns the server info after initialization.

pub fn server_capabilities(&self) -> &ServerCapabilities

Returns the server capabilities after initialization.

pub fn protocol_version(&self) -> &str

Returns the protocol version negotiated during initialization.

pub fn cancel_request( &mut self, request_id: impl Into<RequestId>, reason: Option<String>, await_cleanup: bool, ) -> Result<(), McpError>

Sends a cancellation notification for a previously issued request.

Set await_cleanup to request that the server wait for any cleanup before acknowledging completion (best-effort; server-dependent).

Errors

Returns an error if the notification cannot be sent.

pub fn list_tools(&mut self) -> Result<Vec<Tool>, McpError>

Lists available tools.

Errors

Returns an error if the request fails.

pub fn call_tool( &mut self, name: &str, arguments: Value, ) -> Result<Vec<Content>, McpError>

Calls a tool with the given arguments.

Errors

Returns an error if the tool call fails.

pub fn call_tool_with_progress( &mut self, name: &str, arguments: Value, on_progress: &mut dyn FnMut(f64, Option<f64>, Option<&str>), ) -> Result<Vec<Content>, McpError>

Calls a tool with progress callback support.

This method allows you to receive progress notifications during tool execution. The callback is invoked for each progress notification received from the server.

Arguments

• name - The tool name to call
• arguments - The tool arguments as JSON
• on_progress - Callback invoked for each progress notification

Errors

Returns an error if the tool call fails.

pub fn list_resources(&mut self) -> Result<Vec<Resource>, McpError>

Lists available resources.

Errors

Returns an error if the request fails.

pub fn list_resource_templates( &mut self, ) -> Result<Vec<ResourceTemplate>, McpError>

Lists available resource templates.

Errors

Returns an error if the request fails.

pub fn set_log_level(&mut self, level: LogLevel) -> Result<(), McpError>

Sets the server log level (if supported).

Errors

Returns an error if the request fails.

pub fn read_resource( &mut self, uri: &str, ) -> Result<Vec<ResourceContent>, McpError>

Reads a resource by URI.

Errors

Returns an error if the resource cannot be read.

pub fn list_prompts(&mut self) -> Result<Vec<Prompt>, McpError>

Lists available prompts.

Errors

Returns an error if the request fails.

pub fn get_prompt( &mut self, name: &str, arguments: HashMap<String, String>, ) -> Result<Vec<PromptMessage>, McpError>

Gets a prompt with the given arguments.

Errors

Returns an error if the prompt cannot be retrieved.

pub fn submit_task( &mut self, task_type: &str, input: Value, ) -> Result<TaskInfo, McpError>

Submits a background task for execution.

Arguments

• task_type - The type of task to execute (e.g., “data_export”, “batch_process”)
• input - Task parameters as JSON

Errors

Returns an error if the server doesn’t support tasks or the request fails.

pub fn list_tasks( &mut self, status: Option<TaskStatus>, cursor: Option<&str>, limit: Option<u32>, ) -> Result<ListTasksResult, McpError>

Lists tasks with optional status filter.

Arguments

• status - Optional filter by task status
• cursor - Optional pagination cursor from previous response

Errors

Returns an error if the server doesn’t support tasks or the request fails.

pub fn list_tasks_all( &mut self, status: Option<TaskStatus>, ) -> Result<Vec<TaskInfo>, McpError>

Lists all tasks by following pagination cursors until exhaustion.

Errors

Returns an error if any request fails.

pub fn get_task(&mut self, task_id: &str) -> Result<GetTaskResult, McpError>

Gets detailed information about a specific task.

Arguments

• task_id - The task ID to retrieve

Errors

Returns an error if the task is not found or the request fails.

pub fn cancel_task(&mut self, task_id: &str) -> Result<TaskInfo, McpError>

Cancels a running or pending task.

Arguments

• task_id - The task ID to cancel

Errors

Returns an error if the task cannot be cancelled or is already complete.

pub fn cancel_task_with_reason( &mut self, task_id: &str, reason: Option<&str>, ) -> Result<TaskInfo, McpError>

Cancels a running or pending task with an optional reason.

Arguments

• task_id - The task ID to cancel
• reason - Optional reason for the cancellation

Errors

Returns an error if the task cannot be cancelled or is already complete.

pub fn wait_for_task( &mut self, task_id: &str, poll_interval: Duration, ) -> Result<TaskResult, McpError>

Waits for a task to complete by polling.

This method polls the server at the specified interval until the task reaches a terminal state (completed, failed, or cancelled).

Arguments

• task_id - The task ID to wait for
• poll_interval - Duration between poll requests

Errors

Returns an error if the task fails, is cancelled, or the request fails.

pub fn wait_for_task_with_progress<F>( &mut self, task_id: &str, poll_interval: Duration, on_progress: F, ) -> Result<TaskResult, McpError>

where F: FnMut(f64, Option<&str>),

Waits for a task with progress callback.

Similar to wait_for_task but also provides progress information via callback.

Arguments

• task_id - The task ID to wait for
• poll_interval - Duration between poll requests
• on_progress - Callback invoked with progress updates

Errors

Returns an error if the task fails, is cancelled, or the request fails.

pub fn close(self)

Closes the client connection.

Trait Implementations

impl Drop for Client

fn drop(&mut self)

Executes the destructor for this type.

fn pin_drop(self: Pin<&mut Self>)

🔬This is a nightly-only experimental API. (pin_ergonomics)

Execute the destructor for this type, but different to Drop::drop, it requires self to be pinned.

impl ProxyBackend for Client

fn list_tools(&mut self) -> Result<Vec<Tool>, McpError>

Lists available tools.

fn list_resources(&mut self) -> Result<Vec<Resource>, McpError>

Lists available resources.

fn list_resource_templates(&mut self) -> Result<Vec<ResourceTemplate>, McpError>

Lists available resource templates.

fn list_prompts(&mut self) -> Result<Vec<Prompt>, McpError>

Lists available prompts.

fn call_tool( &mut self, name: &str, arguments: Value, ) -> Result<Vec<Content>, McpError>

Calls a tool.

fn call_tool_with_progress( &mut self, name: &str, arguments: Value, on_progress: &mut dyn FnMut(f64, Option<f64>, Option<String>), ) -> Result<Vec<Content>, McpError>

Calls a tool with progress callback support.

fn read_resource(&mut self, uri: &str) -> Result<Vec<ResourceContent>, McpError>

Reads a resource by URI.

fn get_prompt( &mut self, name: &str, arguments: HashMap<String, String>, ) -> Result<Vec<PromptMessage>, McpError>

Fetches a prompt by name.