@modelcontextprotocol/sdk
Version:
Model Context Protocol implementation for TypeScript
1,422 lines • 69 kB
JavaScript
import * as z from 'zod/v4';
export const LATEST_PROTOCOL_VERSION = '2025-11-25';
export const DEFAULT_NEGOTIATED_PROTOCOL_VERSION = '2025-03-26';
export const SUPPORTED_PROTOCOL_VERSIONS = [LATEST_PROTOCOL_VERSION, '2025-06-18', '2025-03-26', '2024-11-05', '2024-10-07'];
export const RELATED_TASK_META_KEY = 'io.modelcontextprotocol/related-task';
/* JSON-RPC types */
export const JSONRPC_VERSION = '2.0';
/**
* Assert 'object' type schema.
*
* @internal
*/
const AssertObjectSchema = z.custom((v) => v !== null && (typeof v === 'object' || typeof v === 'function'));
/**
* A progress token, used to associate progress notifications with the original request.
*/
export const ProgressTokenSchema = z.union([z.string(), z.number().int()]);
/**
* An opaque token used to represent a cursor for pagination.
*/
export const CursorSchema = z.string();
/**
* Task creation parameters, used to ask that the server create a task to represent a request.
*/
export const TaskCreationParamsSchema = z.looseObject({
/**
* Time in milliseconds to keep task results available after completion.
* If null, the task has unlimited lifetime until manually cleaned up.
*/
ttl: z.union([z.number(), z.null()]).optional(),
/**
* Time in milliseconds to wait between task status requests.
*/
pollInterval: z.number().optional()
});
/**
* Task association metadata, used to signal which task a message originated from.
*/
export const RelatedTaskMetadataSchema = z.looseObject({
taskId: z.string()
});
const RequestMetaSchema = z.looseObject({
/**
* If specified, the caller is requesting out-of-band progress notifications for this request (as represented by notifications/progress). The value of this parameter is an opaque token that will be attached to any subsequent notifications. The receiver is not obligated to provide these notifications.
*/
progressToken: ProgressTokenSchema.optional(),
/**
* If specified, this request is related to the provided task.
*/
[RELATED_TASK_META_KEY]: RelatedTaskMetadataSchema.optional()
});
/**
* Common params for any request.
*/
const BaseRequestParamsSchema = z.looseObject({
/**
* If specified, the caller is requesting that the receiver create a task to represent the request.
* Task creation parameters are now at the top level instead of in _meta.
*/
task: TaskCreationParamsSchema.optional(),
/**
* See [General fields: `_meta`](/specification/draft/basic/index#meta) for notes on `_meta` usage.
*/
_meta: RequestMetaSchema.optional()
});
export const RequestSchema = z.object({
method: z.string(),
params: BaseRequestParamsSchema.optional()
});
const NotificationsParamsSchema = z.looseObject({
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z
.object({
/**
* If specified, this notification is related to the provided task.
*/
[RELATED_TASK_META_KEY]: z.optional(RelatedTaskMetadataSchema)
})
.passthrough()
.optional()
});
export const NotificationSchema = z.object({
method: z.string(),
params: NotificationsParamsSchema.optional()
});
export const ResultSchema = z.looseObject({
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z
.looseObject({
/**
* If specified, this result is related to the provided task.
*/
[RELATED_TASK_META_KEY]: RelatedTaskMetadataSchema.optional()
})
.optional()
});
/**
* A uniquely identifying ID for a request in JSON-RPC.
*/
export const RequestIdSchema = z.union([z.string(), z.number().int()]);
/**
* A request that expects a response.
*/
export const JSONRPCRequestSchema = z
.object({
jsonrpc: z.literal(JSONRPC_VERSION),
id: RequestIdSchema,
...RequestSchema.shape
})
.strict();
export const isJSONRPCRequest = (value) => JSONRPCRequestSchema.safeParse(value).success;
/**
* A notification which does not expect a response.
*/
export const JSONRPCNotificationSchema = z
.object({
jsonrpc: z.literal(JSONRPC_VERSION),
...NotificationSchema.shape
})
.strict();
export const isJSONRPCNotification = (value) => JSONRPCNotificationSchema.safeParse(value).success;
/**
* A successful (non-error) response to a request.
*/
export const JSONRPCResponseSchema = z
.object({
jsonrpc: z.literal(JSONRPC_VERSION),
id: RequestIdSchema,
result: ResultSchema
})
.strict();
export const isJSONRPCResponse = (value) => JSONRPCResponseSchema.safeParse(value).success;
/**
* Error codes defined by the JSON-RPC specification.
*/
export var ErrorCode;
(function (ErrorCode) {
// SDK error codes
ErrorCode[ErrorCode["ConnectionClosed"] = -32000] = "ConnectionClosed";
ErrorCode[ErrorCode["RequestTimeout"] = -32001] = "RequestTimeout";
// Standard JSON-RPC error codes
ErrorCode[ErrorCode["ParseError"] = -32700] = "ParseError";
ErrorCode[ErrorCode["InvalidRequest"] = -32600] = "InvalidRequest";
ErrorCode[ErrorCode["MethodNotFound"] = -32601] = "MethodNotFound";
ErrorCode[ErrorCode["InvalidParams"] = -32602] = "InvalidParams";
ErrorCode[ErrorCode["InternalError"] = -32603] = "InternalError";
// MCP-specific error codes
ErrorCode[ErrorCode["UrlElicitationRequired"] = -32042] = "UrlElicitationRequired";
})(ErrorCode || (ErrorCode = {}));
/**
* A response to a request that indicates an error occurred.
*/
export const JSONRPCErrorSchema = z
.object({
jsonrpc: z.literal(JSONRPC_VERSION),
id: RequestIdSchema,
error: z.object({
/**
* The error type that occurred.
*/
code: z.number().int(),
/**
* A short description of the error. The message SHOULD be limited to a concise single sentence.
*/
message: z.string(),
/**
* Additional information about the error. The value of this member is defined by the sender (e.g. detailed error information, nested errors etc.).
*/
data: z.optional(z.unknown())
})
})
.strict();
export const isJSONRPCError = (value) => JSONRPCErrorSchema.safeParse(value).success;
export const JSONRPCMessageSchema = z.union([JSONRPCRequestSchema, JSONRPCNotificationSchema, JSONRPCResponseSchema, JSONRPCErrorSchema]);
/* Empty result */
/**
* A response that indicates success but carries no data.
*/
export const EmptyResultSchema = ResultSchema.strict();
export const CancelledNotificationParamsSchema = NotificationsParamsSchema.extend({
/**
* The ID of the request to cancel.
*
* This MUST correspond to the ID of a request previously issued in the same direction.
*/
requestId: RequestIdSchema,
/**
* An optional string describing the reason for the cancellation. This MAY be logged or presented to the user.
*/
reason: z.string().optional()
});
/* Cancellation */
/**
* This notification can be sent by either side to indicate that it is cancelling a previously-issued request.
*
* The request SHOULD still be in-flight, but due to communication latency, it is always possible that this notification MAY arrive after the request has already finished.
*
* This notification indicates that the result will be unused, so any associated processing SHOULD cease.
*
* A client MUST NOT attempt to cancel its `initialize` request.
*/
export const CancelledNotificationSchema = NotificationSchema.extend({
method: z.literal('notifications/cancelled'),
params: CancelledNotificationParamsSchema
});
/* Base Metadata */
/**
* Icon schema for use in tools, prompts, resources, and implementations.
*/
export const IconSchema = z.object({
/**
* URL or data URI for the icon.
*/
src: z.string(),
/**
* Optional MIME type for the icon.
*/
mimeType: z.string().optional(),
/**
* Optional array of strings that specify sizes at which the icon can be used.
* Each string should be in WxH format (e.g., `"48x48"`, `"96x96"`) or `"any"` for scalable formats like SVG.
*
* If not provided, the client should assume that the icon can be used at any size.
*/
sizes: z.array(z.string()).optional()
});
/**
* Base schema to add `icons` property.
*
*/
export const IconsSchema = z.object({
/**
* Optional set of sized icons that the client can display in a user interface.
*
* Clients that support rendering icons MUST support at least the following MIME types:
* - `image/png` - PNG images (safe, universal compatibility)
* - `image/jpeg` (and `image/jpg`) - JPEG images (safe, universal compatibility)
*
* Clients that support rendering icons SHOULD also support:
* - `image/svg+xml` - SVG images (scalable but requires security precautions)
* - `image/webp` - WebP images (modern, efficient format)
*/
icons: z.array(IconSchema).optional()
});
/**
* Base metadata interface for common properties across resources, tools, prompts, and implementations.
*/
export const BaseMetadataSchema = z.object({
/** Intended for programmatic or logical use, but used as a display name in past specs or fallback */
name: z.string(),
/**
* Intended for UI and end-user contexts — optimized to be human-readable and easily understood,
* even by those unfamiliar with domain-specific terminology.
*
* If not provided, the name should be used for display (except for Tool,
* where `annotations.title` should be given precedence over using `name`,
* if present).
*/
title: z.string().optional()
});
/* Initialization */
/**
* Describes the name and version of an MCP implementation.
*/
export const ImplementationSchema = BaseMetadataSchema.extend({
...BaseMetadataSchema.shape,
...IconsSchema.shape,
version: z.string(),
/**
* An optional URL of the website for this implementation.
*/
websiteUrl: z.string().optional()
});
const FormElicitationCapabilitySchema = z.intersection(z.object({
applyDefaults: z.boolean().optional()
}), z.record(z.string(), z.unknown()));
const ElicitationCapabilitySchema = z.preprocess(value => {
if (value && typeof value === 'object' && !Array.isArray(value)) {
if (Object.keys(value).length === 0) {
return { form: {} };
}
}
return value;
}, z.intersection(z.object({
form: FormElicitationCapabilitySchema.optional(),
url: AssertObjectSchema.optional()
}), z.record(z.string(), z.unknown()).optional()));
/**
* Task capabilities for clients, indicating which request types support task creation.
*/
export const ClientTasksCapabilitySchema = z
.object({
/**
* Present if the client supports listing tasks.
*/
list: z.optional(z.object({}).passthrough()),
/**
* Present if the client supports cancelling tasks.
*/
cancel: z.optional(z.object({}).passthrough()),
/**
* Capabilities for task creation on specific request types.
*/
requests: z.optional(z
.object({
/**
* Task support for sampling requests.
*/
sampling: z.optional(z
.object({
createMessage: z.optional(z.object({}).passthrough())
})
.passthrough()),
/**
* Task support for elicitation requests.
*/
elicitation: z.optional(z
.object({
create: z.optional(z.object({}).passthrough())
})
.passthrough())
})
.passthrough())
})
.passthrough();
/**
* Task capabilities for servers, indicating which request types support task creation.
*/
export const ServerTasksCapabilitySchema = z
.object({
/**
* Present if the server supports listing tasks.
*/
list: z.optional(z.object({}).passthrough()),
/**
* Present if the server supports cancelling tasks.
*/
cancel: z.optional(z.object({}).passthrough()),
/**
* Capabilities for task creation on specific request types.
*/
requests: z.optional(z
.object({
/**
* Task support for tool requests.
*/
tools: z.optional(z
.object({
call: z.optional(z.object({}).passthrough())
})
.passthrough())
})
.passthrough())
})
.passthrough();
/**
* Capabilities a client may support. Known capabilities are defined here, in this schema, but this is not a closed set: any client can define its own, additional capabilities.
*/
export const ClientCapabilitiesSchema = z.object({
/**
* Experimental, non-standard capabilities that the client supports.
*/
experimental: z.record(z.string(), AssertObjectSchema).optional(),
/**
* Present if the client supports sampling from an LLM.
*/
sampling: z
.object({
/**
* Present if the client supports context inclusion via includeContext parameter.
* If not declared, servers SHOULD only use `includeContext: "none"` (or omit it).
*/
context: AssertObjectSchema.optional(),
/**
* Present if the client supports tool use via tools and toolChoice parameters.
*/
tools: AssertObjectSchema.optional()
})
.optional(),
/**
* Present if the client supports eliciting user input.
*/
elicitation: ElicitationCapabilitySchema.optional(),
/**
* Present if the client supports listing roots.
*/
roots: z
.object({
/**
* Whether the client supports issuing notifications for changes to the roots list.
*/
listChanged: z.boolean().optional()
})
.optional(),
/**
* Present if the client supports task creation.
*/
tasks: z.optional(ClientTasksCapabilitySchema)
});
export const InitializeRequestParamsSchema = BaseRequestParamsSchema.extend({
/**
* The latest version of the Model Context Protocol that the client supports. The client MAY decide to support older versions as well.
*/
protocolVersion: z.string(),
capabilities: ClientCapabilitiesSchema,
clientInfo: ImplementationSchema
});
/**
* This request is sent from the client to the server when it first connects, asking it to begin initialization.
*/
export const InitializeRequestSchema = RequestSchema.extend({
method: z.literal('initialize'),
params: InitializeRequestParamsSchema
});
export const isInitializeRequest = (value) => InitializeRequestSchema.safeParse(value).success;
/**
* Capabilities that a server may support. Known capabilities are defined here, in this schema, but this is not a closed set: any server can define its own, additional capabilities.
*/
export const ServerCapabilitiesSchema = z
.object({
/**
* Experimental, non-standard capabilities that the server supports.
*/
experimental: z.record(z.string(), AssertObjectSchema).optional(),
/**
* Present if the server supports sending log messages to the client.
*/
logging: AssertObjectSchema.optional(),
/**
* Present if the server supports sending completions to the client.
*/
completions: AssertObjectSchema.optional(),
/**
* Present if the server offers any prompt templates.
*/
prompts: z.optional(z.object({
/**
* Whether this server supports issuing notifications for changes to the prompt list.
*/
listChanged: z.optional(z.boolean())
})),
/**
* Present if the server offers any resources to read.
*/
resources: z
.object({
/**
* Whether this server supports clients subscribing to resource updates.
*/
subscribe: z.boolean().optional(),
/**
* Whether this server supports issuing notifications for changes to the resource list.
*/
listChanged: z.boolean().optional()
})
.optional(),
/**
* Present if the server offers any tools to call.
*/
tools: z
.object({
/**
* Whether this server supports issuing notifications for changes to the tool list.
*/
listChanged: z.boolean().optional()
})
.optional(),
/**
* Present if the server supports task creation.
*/
tasks: z.optional(ServerTasksCapabilitySchema)
})
.passthrough();
/**
* After receiving an initialize request from the client, the server sends this response.
*/
export const InitializeResultSchema = ResultSchema.extend({
/**
* The version of the Model Context Protocol that the server wants to use. This may not match the version that the client requested. If the client cannot support this version, it MUST disconnect.
*/
protocolVersion: z.string(),
capabilities: ServerCapabilitiesSchema,
serverInfo: ImplementationSchema,
/**
* Instructions describing how to use the server and its features.
*
* This can be used by clients to improve the LLM's understanding of available tools, resources, etc. It can be thought of like a "hint" to the model. For example, this information MAY be added to the system prompt.
*/
instructions: z.string().optional()
});
/**
* This notification is sent from the client to the server after initialization has finished.
*/
export const InitializedNotificationSchema = NotificationSchema.extend({
method: z.literal('notifications/initialized')
});
export const isInitializedNotification = (value) => InitializedNotificationSchema.safeParse(value).success;
/* Ping */
/**
* A ping, issued by either the server or the client, to check that the other party is still alive. The receiver must promptly respond, or else may be disconnected.
*/
export const PingRequestSchema = RequestSchema.extend({
method: z.literal('ping')
});
/* Progress notifications */
export const ProgressSchema = z.object({
/**
* The progress thus far. This should increase every time progress is made, even if the total is unknown.
*/
progress: z.number(),
/**
* Total number of items to process (or total progress required), if known.
*/
total: z.optional(z.number()),
/**
* An optional message describing the current progress.
*/
message: z.optional(z.string())
});
export const ProgressNotificationParamsSchema = z.object({
...NotificationsParamsSchema.shape,
...ProgressSchema.shape,
/**
* The progress token which was given in the initial request, used to associate this notification with the request that is proceeding.
*/
progressToken: ProgressTokenSchema
});
/**
* An out-of-band notification used to inform the receiver of a progress update for a long-running request.
*
* @category notifications/progress
*/
export const ProgressNotificationSchema = NotificationSchema.extend({
method: z.literal('notifications/progress'),
params: ProgressNotificationParamsSchema
});
export const PaginatedRequestParamsSchema = BaseRequestParamsSchema.extend({
/**
* An opaque token representing the current pagination position.
* If provided, the server should return results starting after this cursor.
*/
cursor: CursorSchema.optional()
});
/* Pagination */
export const PaginatedRequestSchema = RequestSchema.extend({
params: PaginatedRequestParamsSchema.optional()
});
export const PaginatedResultSchema = ResultSchema.extend({
/**
* An opaque token representing the pagination position after the last returned result.
* If present, there may be more results available.
*/
nextCursor: z.optional(CursorSchema)
});
/* Tasks */
/**
* A pollable state object associated with a request.
*/
export const TaskSchema = z.object({
taskId: z.string(),
status: z.enum(['working', 'input_required', 'completed', 'failed', 'cancelled']),
/**
* Time in milliseconds to keep task results available after completion.
* If null, the task has unlimited lifetime until manually cleaned up.
*/
ttl: z.union([z.number(), z.null()]),
/**
* ISO 8601 timestamp when the task was created.
*/
createdAt: z.string(),
/**
* ISO 8601 timestamp when the task was last updated.
*/
lastUpdatedAt: z.string(),
pollInterval: z.optional(z.number()),
/**
* Optional diagnostic message for failed tasks or other status information.
*/
statusMessage: z.optional(z.string())
});
/**
* Result returned when a task is created, containing the task data wrapped in a task field.
*/
export const CreateTaskResultSchema = ResultSchema.extend({
task: TaskSchema
});
/**
* Parameters for task status notification.
*/
export const TaskStatusNotificationParamsSchema = NotificationsParamsSchema.merge(TaskSchema);
/**
* A notification sent when a task's status changes.
*/
export const TaskStatusNotificationSchema = NotificationSchema.extend({
method: z.literal('notifications/tasks/status'),
params: TaskStatusNotificationParamsSchema
});
/**
* A request to get the state of a specific task.
*/
export const GetTaskRequestSchema = RequestSchema.extend({
method: z.literal('tasks/get'),
params: BaseRequestParamsSchema.extend({
taskId: z.string()
})
});
/**
* The response to a tasks/get request.
*/
export const GetTaskResultSchema = ResultSchema.merge(TaskSchema);
/**
* A request to get the result of a specific task.
*/
export const GetTaskPayloadRequestSchema = RequestSchema.extend({
method: z.literal('tasks/result'),
params: BaseRequestParamsSchema.extend({
taskId: z.string()
})
});
/**
* A request to list tasks.
*/
export const ListTasksRequestSchema = PaginatedRequestSchema.extend({
method: z.literal('tasks/list')
});
/**
* The response to a tasks/list request.
*/
export const ListTasksResultSchema = PaginatedResultSchema.extend({
tasks: z.array(TaskSchema)
});
/**
* A request to cancel a specific task.
*/
export const CancelTaskRequestSchema = RequestSchema.extend({
method: z.literal('tasks/cancel'),
params: BaseRequestParamsSchema.extend({
taskId: z.string()
})
});
/**
* The response to a tasks/cancel request.
*/
export const CancelTaskResultSchema = ResultSchema.merge(TaskSchema);
/* Resources */
/**
* The contents of a specific resource or sub-resource.
*/
export const ResourceContentsSchema = z.object({
/**
* The URI of this resource.
*/
uri: z.string(),
/**
* The MIME type of this resource, if known.
*/
mimeType: z.optional(z.string()),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.record(z.string(), z.unknown()).optional()
});
export const TextResourceContentsSchema = ResourceContentsSchema.extend({
/**
* The text of the item. This must only be set if the item can actually be represented as text (not binary data).
*/
text: z.string()
});
/**
* A Zod schema for validating Base64 strings that is more performant and
* robust for very large inputs than the default regex-based check. It avoids
* stack overflows by using the native `atob` function for validation.
*/
const Base64Schema = z.string().refine(val => {
try {
// atob throws a DOMException if the string contains characters
// that are not part of the Base64 character set.
atob(val);
return true;
}
catch (_a) {
return false;
}
}, { message: 'Invalid Base64 string' });
export const BlobResourceContentsSchema = ResourceContentsSchema.extend({
/**
* A base64-encoded string representing the binary data of the item.
*/
blob: Base64Schema
});
/**
* Optional annotations providing clients additional context about a resource.
*/
export const AnnotationsSchema = z.object({
/**
* Intended audience(s) for the resource.
*/
audience: z.array(z.enum(['user', 'assistant'])).optional(),
/**
* Importance hint for the resource, from 0 (least) to 1 (most).
*/
priority: z.number().min(0).max(1).optional(),
/**
* ISO 8601 timestamp for the most recent modification.
*/
lastModified: z.iso.datetime({ offset: true }).optional()
});
/**
* A known resource that the server is capable of reading.
*/
export const ResourceSchema = z.object({
...BaseMetadataSchema.shape,
...IconsSchema.shape,
/**
* The URI of this resource.
*/
uri: z.string(),
/**
* A description of what this resource represents.
*
* This can be used by clients to improve the LLM's understanding of available resources. It can be thought of like a "hint" to the model.
*/
description: z.optional(z.string()),
/**
* The MIME type of this resource, if known.
*/
mimeType: z.optional(z.string()),
/**
* Optional annotations for the client.
*/
annotations: AnnotationsSchema.optional(),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.optional(z.looseObject({}))
});
/**
* A template description for resources available on the server.
*/
export const ResourceTemplateSchema = z.object({
...BaseMetadataSchema.shape,
...IconsSchema.shape,
/**
* A URI template (according to RFC 6570) that can be used to construct resource URIs.
*/
uriTemplate: z.string(),
/**
* A description of what this template is for.
*
* This can be used by clients to improve the LLM's understanding of available resources. It can be thought of like a "hint" to the model.
*/
description: z.optional(z.string()),
/**
* The MIME type for all resources that match this template. This should only be included if all resources matching this template have the same type.
*/
mimeType: z.optional(z.string()),
/**
* Optional annotations for the client.
*/
annotations: AnnotationsSchema.optional(),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.optional(z.looseObject({}))
});
/**
* Sent from the client to request a list of resources the server has.
*/
export const ListResourcesRequestSchema = PaginatedRequestSchema.extend({
method: z.literal('resources/list')
});
/**
* The server's response to a resources/list request from the client.
*/
export const ListResourcesResultSchema = PaginatedResultSchema.extend({
resources: z.array(ResourceSchema)
});
/**
* Sent from the client to request a list of resource templates the server has.
*/
export const ListResourceTemplatesRequestSchema = PaginatedRequestSchema.extend({
method: z.literal('resources/templates/list')
});
/**
* The server's response to a resources/templates/list request from the client.
*/
export const ListResourceTemplatesResultSchema = PaginatedResultSchema.extend({
resourceTemplates: z.array(ResourceTemplateSchema)
});
export const ResourceRequestParamsSchema = BaseRequestParamsSchema.extend({
/**
* The URI of the resource to read. The URI can use any protocol; it is up to the server how to interpret it.
*
* @format uri
*/
uri: z.string()
});
/**
* Parameters for a `resources/read` request.
*/
export const ReadResourceRequestParamsSchema = ResourceRequestParamsSchema;
/**
* Sent from the client to the server, to read a specific resource URI.
*/
export const ReadResourceRequestSchema = RequestSchema.extend({
method: z.literal('resources/read'),
params: ReadResourceRequestParamsSchema
});
/**
* The server's response to a resources/read request from the client.
*/
export const ReadResourceResultSchema = ResultSchema.extend({
contents: z.array(z.union([TextResourceContentsSchema, BlobResourceContentsSchema]))
});
/**
* An optional notification from the server to the client, informing it that the list of resources it can read from has changed. This may be issued by servers without any previous subscription from the client.
*/
export const ResourceListChangedNotificationSchema = NotificationSchema.extend({
method: z.literal('notifications/resources/list_changed')
});
export const SubscribeRequestParamsSchema = ResourceRequestParamsSchema;
/**
* Sent from the client to request resources/updated notifications from the server whenever a particular resource changes.
*/
export const SubscribeRequestSchema = RequestSchema.extend({
method: z.literal('resources/subscribe'),
params: SubscribeRequestParamsSchema
});
export const UnsubscribeRequestParamsSchema = ResourceRequestParamsSchema;
/**
* Sent from the client to request cancellation of resources/updated notifications from the server. This should follow a previous resources/subscribe request.
*/
export const UnsubscribeRequestSchema = RequestSchema.extend({
method: z.literal('resources/unsubscribe'),
params: UnsubscribeRequestParamsSchema
});
/**
* Parameters for a `notifications/resources/updated` notification.
*/
export const ResourceUpdatedNotificationParamsSchema = NotificationsParamsSchema.extend({
/**
* The URI of the resource that has been updated. This might be a sub-resource of the one that the client actually subscribed to.
*/
uri: z.string()
});
/**
* A notification from the server to the client, informing it that a resource has changed and may need to be read again. This should only be sent if the client previously sent a resources/subscribe request.
*/
export const ResourceUpdatedNotificationSchema = NotificationSchema.extend({
method: z.literal('notifications/resources/updated'),
params: ResourceUpdatedNotificationParamsSchema
});
/* Prompts */
/**
* Describes an argument that a prompt can accept.
*/
export const PromptArgumentSchema = z.object({
/**
* The name of the argument.
*/
name: z.string(),
/**
* A human-readable description of the argument.
*/
description: z.optional(z.string()),
/**
* Whether this argument must be provided.
*/
required: z.optional(z.boolean())
});
/**
* A prompt or prompt template that the server offers.
*/
export const PromptSchema = z.object({
...BaseMetadataSchema.shape,
...IconsSchema.shape,
/**
* An optional description of what this prompt provides
*/
description: z.optional(z.string()),
/**
* A list of arguments to use for templating the prompt.
*/
arguments: z.optional(z.array(PromptArgumentSchema)),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.optional(z.looseObject({}))
});
/**
* Sent from the client to request a list of prompts and prompt templates the server has.
*/
export const ListPromptsRequestSchema = PaginatedRequestSchema.extend({
method: z.literal('prompts/list')
});
/**
* The server's response to a prompts/list request from the client.
*/
export const ListPromptsResultSchema = PaginatedResultSchema.extend({
prompts: z.array(PromptSchema)
});
/**
* Parameters for a `prompts/get` request.
*/
export const GetPromptRequestParamsSchema = BaseRequestParamsSchema.extend({
/**
* The name of the prompt or prompt template.
*/
name: z.string(),
/**
* Arguments to use for templating the prompt.
*/
arguments: z.record(z.string(), z.string()).optional()
});
/**
* Used by the client to get a prompt provided by the server.
*/
export const GetPromptRequestSchema = RequestSchema.extend({
method: z.literal('prompts/get'),
params: GetPromptRequestParamsSchema
});
/**
* Text provided to or from an LLM.
*/
export const TextContentSchema = z.object({
type: z.literal('text'),
/**
* The text content of the message.
*/
text: z.string(),
/**
* Optional annotations for the client.
*/
annotations: AnnotationsSchema.optional(),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.record(z.string(), z.unknown()).optional()
});
/**
* An image provided to or from an LLM.
*/
export const ImageContentSchema = z.object({
type: z.literal('image'),
/**
* The base64-encoded image data.
*/
data: Base64Schema,
/**
* The MIME type of the image. Different providers may support different image types.
*/
mimeType: z.string(),
/**
* Optional annotations for the client.
*/
annotations: AnnotationsSchema.optional(),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.record(z.string(), z.unknown()).optional()
});
/**
* An Audio provided to or from an LLM.
*/
export const AudioContentSchema = z.object({
type: z.literal('audio'),
/**
* The base64-encoded audio data.
*/
data: Base64Schema,
/**
* The MIME type of the audio. Different providers may support different audio types.
*/
mimeType: z.string(),
/**
* Optional annotations for the client.
*/
annotations: AnnotationsSchema.optional(),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.record(z.string(), z.unknown()).optional()
});
/**
* A tool call request from an assistant (LLM).
* Represents the assistant's request to use a tool.
*/
export const ToolUseContentSchema = z
.object({
type: z.literal('tool_use'),
/**
* The name of the tool to invoke.
* Must match a tool name from the request's tools array.
*/
name: z.string(),
/**
* Unique identifier for this tool call.
* Used to correlate with ToolResultContent in subsequent messages.
*/
id: z.string(),
/**
* Arguments to pass to the tool.
* Must conform to the tool's inputSchema.
*/
input: z.object({}).passthrough(),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.optional(z.object({}).passthrough())
})
.passthrough();
/**
* The contents of a resource, embedded into a prompt or tool call result.
*/
export const EmbeddedResourceSchema = z.object({
type: z.literal('resource'),
resource: z.union([TextResourceContentsSchema, BlobResourceContentsSchema]),
/**
* Optional annotations for the client.
*/
annotations: AnnotationsSchema.optional(),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.record(z.string(), z.unknown()).optional()
});
/**
* A resource that the server is capable of reading, included in a prompt or tool call result.
*
* Note: resource links returned by tools are not guaranteed to appear in the results of `resources/list` requests.
*/
export const ResourceLinkSchema = ResourceSchema.extend({
type: z.literal('resource_link')
});
/**
* A content block that can be used in prompts and tool results.
*/
export const ContentBlockSchema = z.union([
TextContentSchema,
ImageContentSchema,
AudioContentSchema,
ResourceLinkSchema,
EmbeddedResourceSchema
]);
/**
* Describes a message returned as part of a prompt.
*/
export const PromptMessageSchema = z.object({
role: z.enum(['user', 'assistant']),
content: ContentBlockSchema
});
/**
* The server's response to a prompts/get request from the client.
*/
export const GetPromptResultSchema = ResultSchema.extend({
/**
* An optional description for the prompt.
*/
description: z.optional(z.string()),
messages: z.array(PromptMessageSchema)
});
/**
* An optional notification from the server to the client, informing it that the list of prompts it offers has changed. This may be issued by servers without any previous subscription from the client.
*/
export const PromptListChangedNotificationSchema = NotificationSchema.extend({
method: z.literal('notifications/prompts/list_changed')
});
/* Tools */
/**
* Additional properties describing a Tool to clients.
*
* NOTE: all properties in ToolAnnotations are **hints**.
* They are not guaranteed to provide a faithful description of
* tool behavior (including descriptive properties like `title`).
*
* Clients should never make tool use decisions based on ToolAnnotations
* received from untrusted servers.
*/
export const ToolAnnotationsSchema = z.object({
/**
* A human-readable title for the tool.
*/
title: z.string().optional(),
/**
* If true, the tool does not modify its environment.
*
* Default: false
*/
readOnlyHint: z.boolean().optional(),
/**
* If true, the tool may perform destructive updates to its environment.
* If false, the tool performs only additive updates.
*
* (This property is meaningful only when `readOnlyHint == false`)
*
* Default: true
*/
destructiveHint: z.boolean().optional(),
/**
* If true, calling the tool repeatedly with the same arguments
* will have no additional effect on the its environment.
*
* (This property is meaningful only when `readOnlyHint == false`)
*
* Default: false
*/
idempotentHint: z.boolean().optional(),
/**
* If true, this tool may interact with an "open world" of external
* entities. If false, the tool's domain of interaction is closed.
* For example, the world of a web search tool is open, whereas that
* of a memory tool is not.
*
* Default: true
*/
openWorldHint: z.boolean().optional()
});
/**
* Execution-related properties for a tool.
*/
export const ToolExecutionSchema = z.object({
/**
* Indicates the tool's preference for task-augmented execution.
* - "required": Clients MUST invoke the tool as a task
* - "optional": Clients MAY invoke the tool as a task or normal request
* - "forbidden": Clients MUST NOT attempt to invoke the tool as a task
*
* If not present, defaults to "forbidden".
*/
taskSupport: z.enum(['required', 'optional', 'forbidden']).optional()
});
/**
* Definition for a tool the client can call.
*/
export const ToolSchema = z.object({
...BaseMetadataSchema.shape,
...IconsSchema.shape,
/**
* A human-readable description of the tool.
*/
description: z.string().optional(),
/**
* A JSON Schema 2020-12 object defining the expected parameters for the tool.
* Must have type: 'object' at the root level per MCP spec.
*/
inputSchema: z
.object({
type: z.literal('object'),
properties: z.record(z.string(), AssertObjectSchema).optional(),
required: z.array(z.string()).optional()
})
.catchall(z.unknown()),
/**
* An optional JSON Schema 2020-12 object defining the structure of the tool's output
* returned in the structuredContent field of a CallToolResult.
* Must have type: 'object' at the root level per MCP spec.
*/
outputSchema: z
.object({
type: z.literal('object'),
properties: z.record(z.string(), AssertObjectSchema).optional(),
required: z.array(z.string()).optional()
})
.catchall(z.unknown())
.optional(),
/**
* Optional additional tool information.
*/
annotations: z.optional(ToolAnnotationsSchema),
/**
* Execution-related properties for this tool.
*/
execution: z.optional(ToolExecutionSchema),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.record(z.string(), z.unknown()).optional()
});
/**
* Sent from the client to request a list of tools the server has.
*/
export const ListToolsRequestSchema = PaginatedRequestSchema.extend({
method: z.literal('tools/list')
});
/**
* The server's response to a tools/list request from the client.
*/
export const ListToolsResultSchema = PaginatedResultSchema.extend({
tools: z.array(ToolSchema)
});
/**
* The server's response to a tool call.
*/
export const CallToolResultSchema = ResultSchema.extend({
/**
* A list of content objects that represent the result of the tool call.
*
* If the Tool does not define an outputSchema, this field MUST be present in the result.
* For backwards compatibility, this field is always present, but it may be empty.
*/
content: z.array(ContentBlockSchema).default([]),
/**
* An object containing structured tool output.
*
* If the Tool defines an outputSchema, this field MUST be present in the result, and contain a JSON object that matches the schema.
*/
structuredContent: z.record(z.string(), z.unknown()).optional(),
/**
* Whether the tool call ended in an error.
*
* If not set, this is assumed to be false (the call was successful).
*
* Any errors that originate from the tool SHOULD be reported inside the result
* object, with `isError` set to true, _not_ as an MCP protocol-level error
* response. Otherwise, the LLM would not be able to see that an error occurred
* and self-correct.
*
* However, any errors in _finding_ the tool, an error indicating that the
* server does not support tool calls, or any other exceptional conditions,
* should be reported as an MCP error response.
*/
isError: z.optional(z.boolean())
});
/**
* CallToolResultSchema extended with backwards compatibility to protocol version 2024-10-07.
*/
export const CompatibilityCallToolResultSchema = CallToolResultSchema.or(ResultSchema.extend({
toolResult: z.unknown()
}));
/**
* Parameters for a `tools/call` request.
*/
export const CallToolRequestParamsSchema = BaseRequestParamsSchema.extend({
/**
* The name of the tool to call.
*/
name: z.string(),
/**
* Arguments to pass to the tool.
*/
arguments: z.optional(z.record(z.string(), z.unknown()))
});
/**
* Used by the client to invoke a tool provided by the server.
*/
export const CallToolRequestSchema = RequestSchema.extend({
method: z.literal('tools/call'),
params: CallToolRequestParamsSchema
});
/**
* An optional notification from the server to the client, informing it that the list of tools it offers has changed. This may be issued by servers without any previous subscription from the client.
*/
export const ToolListChangedNotificationSchema = NotificationSchema.extend({
method: z.literal('notifications/tools/list_changed')
});
/* Logging */
/**
* The severity of a log message.
*/
export const LoggingLevelSchema = z.enum(['debug', 'info', 'notice', 'warning', 'error', 'critical', 'alert', 'emergency']);
/**
* Parameters for a `logging/setLevel` request.
*/
export const SetLevelRequestParamsSchema = BaseRequestParamsSchema.extend({
/**
* The level of logging that the client wants to receive from the server. The server should send all logs at this level and higher (i.e., more severe) to the client as notifications/logging/message.
*/
level: LoggingLevelSchema
});
/**
* A request from the client to the server, to enable or adjust logging.
*/
export const SetLevelRequestSchema = RequestSchema.extend({
method: z.literal('logging/setLevel'),
params: SetLevelRequestParamsSchema
});
/**
* Parameters for a `notifications/message` notification.
*/
export const LoggingMessageNotificationParamsSchema = NotificationsParamsSchema.extend({
/**
* The severity of this log message.
*/
level: LoggingLevelSchema,
/**
* An optional name of the logger issuing this message.
*/
logger: z.string().optional(),
/**
* The data to be logged, such as a string message or an object. Any JSON serializable type is allowed here.
*/
data: z.unknown()
});
/**
* Notification of a log message passed from server to client. If no logging/setLevel request has been sent from the client, the server MAY decide which messages to send automatically.
*/
export const LoggingMessageNotificationSchema = NotificationSchema.extend({
method: z.literal('notifications/message'),
params: LoggingMessageNotificationParamsSchema
});
/* Sampling */
/**
* Hints to use for model selection.
*/
export const ModelHintSchema = z.object({
/**
* A hint for a model name.
*/
name: z.string().optional()
});
/**
* The server's preferences for model selection, requested of the client during sampling.
*/
export const ModelPreferencesSchema = z.object({
/**
* Optional hints to use for model selection.
*/
hints: z.optional(z.array(ModelHintSchema)),
/**
* How much to prioritize cost when selecting a model.
*/
costPriority: z.optional(z.number().min(0).max(1)),
/**
* How much to prioritize sampling speed (latency) when selecting a model.
*/
speedPriority: z.optional(z.number().min(0).max(1)),
/**
* How much to prioritize intelligence and capabilities when selecting a model.
*/
intelligencePriority: z.optional(z.number().min(0).max(1))
});
/**
* Controls tool usage behavior in sampling requests.
*/
export const ToolChoiceSchema = z.object({
/**
* Controls when tools are used:
* - "auto": Model decides whether to use tools (default)
* - "required": Model MUST use at least one tool before completing
* - "none": Model MUST NOT use any tools
*/
mode: z.optional(z.enum(['auto', 'required', 'none']))
});
/**
* The result of a tool execution, provided by the user (server).
* Represents the outcome of invoking a tool requested via ToolUseContent.
*/
export const ToolResultContentSchema = z
.object({
type: z.literal('tool_result'),
toolUseId: z.string().describe('The unique identifier for the corresponding tool call.'),
content: z.array(ContentBlockSchema).default([]),
structuredContent: z.object({}).passthrough().optional(),
isError: z.optional(z.boolean()),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.optional(z.object({}).passthrough())
})
.passthrough();
/**
* Basic content types for sampling responses (without tool use).
* Used for backwards-compatible CreateMessageResult when tools are not used.
*/
export const SamplingContentSchema = z.discriminatedUnion('type', [TextContentSchema, ImageContentSchema, AudioContentSchema]);
/**
* Content block types allowed in sampling messages.
* This includes text, image, audio, tool use requests, and tool results.
*/
export const SamplingMessageContentBlockSchema = z.discriminatedUnion('type', [
TextContentSchema,
ImageContentSchema,
AudioContentSchema,
ToolUseContentSchema,
ToolResultContentSchema
]);
/**
* Describes a message issued to or received from an LLM API.
*/
export const SamplingMessageSchema = z
.object({
role: z.enum(['user', 'assistant']),
content: z.union([SamplingMessageContentBlockSchema, z.array(SamplingMessageContentBlockSchema)]),
/**
* See [MCP specification](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/47339c03c143bb4ec01a26e721a1b8fe66634ebe/docs/specification/draft/basic/index.mdx#general-fields)
* for notes on _meta usage.
*/
_meta: z.optional(z.object({}).passthrough())
})
.passthrough();
/**
* Parameters for a `sampling/createMessage` request.
*/
export const CreateMessageRequestParamsSchema = BaseRequestParamsSchema.extend({
messages: z.array(SamplingMessageSchema),
/**
* The server's preferences for which model to select. The client MAY modify or omit this request.
*/
modelPreferences: ModelPreferencesSchema.optional(),
/**
* An optional system prompt the server wants to use for sampling. The client MAY modify or omit this prompt.
*/
systemPrompt: z.string().optional(),
/**
* A request to include context from one or more MCP servers (including the ca