models Package

An agent implementing the A2A protocol.

An A2A tool stored in a toolbox.

Configuration specific to the A2A protocol.

An agent implementing the A2A protocol.

An A2A tool stored in a toolbox.

A AI Search Index resource.

Configuration specific to the activity protocol.

AgentBlueprintReference.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ManagedAgentIdentityBlueprintReference

AgentCard.

AgentCardSkill.

Insights on set of Agent Evaluation Results.

Insights from the agent cluster analysis.

Agent source for data generation jobs — references an agent to fetch instructions and metadata from.

AgentDefinition.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ExternalAgentDefinition, HostedAgentDefinition, PromptAgentDefinition, WorkflowAgentDefinition

AgentDetails.

AgentEndpointAuthorizationScheme.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: BotServiceAuthorizationScheme, BotServiceRbacAuthorizationScheme, BotServiceTenantAuthorizationScheme, EntraAuthorizationScheme

AgentEndpointConfig.

Agent source for evaluator generation jobs — references an agent to fetch instructions and metadata from.

AgentIdentity.

A persisted issue discovered from an agent's traces.

Additional insight details. Omitted unless details are requested.

Estimated Agent Insights cost.

A highlighted trace that provides evidence for an agent insight.

A lightweight trace reference linked to an agent insight as supporting evidence.

A per-agent Agent Insights monitor that owns configuration, runs, and discovered insights.

Fields accepted when creating an Agent Insights monitor for an agent.

An Agent Insights monitor summary returned by list operations.

Fields that can be updated on an Agent Insights monitor.

A recommended fix for an agent insight.

A customer-renderable change in a proposed fix.

The recommended remediation for an agent insight.

A long-running run that analyzes one agent's traces and updates that agent's insights.

Inputs used when creating an agent insight run.

Custom LROPoller for Agent Insights run operations.

Result statistics produced when an agent insight run succeeds.

Structured reason why scheduled generation is suspended for a monitor.

Token usage for an Agent Insights run.

Fields that can be updated on an agent insight.

The effective overview for an Agent Insights monitor.

A user-provided overview that becomes effective immediately and seeds the next generation.

AgentObjectVersions.

Aggregated evaluation result for a single candidate agent configuration across all tasks.

Evaluation criterion: a name + instruction pair used for per-item scoring.

Base discriminated model for dataset input. Either inline items or a registered reference.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: AgentOptimizationInlineDatasetInput, AgentOptimizationReferenceDatasetInput

A single item in an inline dataset.

Reference to a named evaluator, optionally pinned to a version.

Inline dataset — items supplied directly in the request body.

Agent optimization job resource — a long-running job that optimizes an agent's configuration (instructions, model, skills, tools) to maximize evaluation scores. On success, the result contains scored candidates.

Caller-supplied inputs for an optimization job.

Slim job representation returned by the LIST endpoint.

In-flight progress; only populated while status is queued or in_progress.

Terminal-state result body. Populated when status is succeeded or failed.

Custom LROPoller for agent optimization job operations.

Tuning knobs and run-mode for an optimization job.

Reference to a registered Foundry dataset.

An agent session providing a long-lived compute sandbox for hosted agent invocations.

Input configuration for the evaluation taxonomy when the input type is agent.

AgentVersionDetails.

Agentic identity credential definition.

ApiError.

Error response for API failures.

API Key Credential definition.

Apply patch tool.

ApproximateLocation.

Artifact profile of the model.

Custom AsyncLROPoller for Agent Insights run operations.

Custom AsyncLROPoller for agent optimization job operations.

Custom AsyncLROPoller for data generation job operations.

Custom AsyncLROPoller for evaluator generation job operations.

Custom AsyncLROPoller for Memory Store update operations.

Automatic Code Interpreter Tool Parameters.

Represents a target specifying an Azure AI agent.

Represents a target specifying an Azure AI agent.

Represents a data source for benchmark evaluation runs.

AzureAIDataSourceConfig.

Represents a target specifying an Azure AI model for operations requiring model selection.

Represents a target specifying an Azure AI model for operations requiring model selection.

Represents a data source for evaluation runs that are specific to Continuous Evaluation scenarios.

Azure AI Search Index Definition.

The input definition information for an Azure AI search tool as used to configure an agent.

A set of index resources used by the azure_ai_search tool.

An Azure AI Search tool stored in a toolbox.

The structure for keeping storage queue name and URI.

The definition of Azure function.

AzureFunctionDefinitionFunction.

The structure for keeping storage queue name and URI.

The input definition information for an Azure Function Tool, as used to configure an Agent.

Azure OpenAI model configuration. The API version would be selected by the service for querying the model.

A base class for connection credentials.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: EntraIDCredentials, AgenticIdentityPreviewCredentials, ApiKeyCredentials, CustomCredential, NoAuthenticationCredentials, SASCredentials

A bing custom search configuration.

The input definition information for a Bing custom search tool as used to configure an agent.

The bing custom search tool parameters.

Search configuration for Bing Grounding.

The bing grounding search tool parameters.

The input definition information for a bing grounding search tool as used to configure an agent.

Blob reference details.

SAS Credential definition.

BotServiceAuthorizationScheme.

BotServiceRbacAuthorizationScheme.

BotServiceTenantAuthorizationScheme.

The input definition information for a Browser Automation Tool, as used to configure an Agent.

A browser automation tool stored in a toolbox.

Definition of input parameters for the connection used by the Browser Automation Tool.

Definition of input parameters for the Browser Automation Tool.

A tool for capturing structured outputs.

Coordinates for the analysis chart.

A memory item containing a summary extracted from conversations.

Insights from the cluster analysis.

Token usage for cluster analysis.

Code-based evaluator definition using python code.

Code-based deployment configuration for a hosted agent.

Code interpreter.

A code interpreter tool stored in a toolbox.

Comparison Filter.

Compound Filter.

Computer.

Computer use preview.

Response from the list and get connections operations.

ContainerAutoParam.

Container-based deployment configuration for a hosted agent.

ContainerNetworkPolicyAllowlistParam.

ContainerNetworkPolicyDisabledParam.

ContainerNetworkPolicyDomainSecretParam.

Network access policy for the container.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ContainerNetworkPolicyAllowlistParam, ContainerNetworkPolicyDisabledParam

ContainerSkill.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: InlineSkillParam, SkillReferenceParam

Evaluation rule action for continuous evaluation.

CosmosDB Vector Store Index Definition.

CreateAsyncResponse.

Multipart request body for creating a skill version from files. Accepts either a single zip file or multiple individual skill files (directory upload). For zip uploads, the server extracts and validates contents. For directory uploads, files are validated as-is.

Cron based trigger.

Custom credential definition.

Grammar format.

A custom event routine trigger.

Text format.

Custom tool.

The input format for the custom tool. Default is unconstrained text.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: CustomGrammarFormatParam, CustomTextFormatParam

Daily recurrence schedule.

Data Generation Job resource.

Caller-supplied inputs for a data generation job.

Options for managing data generation jobs.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: SimpleQnADataGenerationJobOptions, SimulationSeedDataGenerationJobOptions, ToolUseFineTuningDataGenerationJobOptions, TracesDataGenerationJobOptions

Output information for a data generation job.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: DatasetDataGenerationJobOutput, FileDataGenerationJobOutput

Output options for data generation job.

Result produced by a successful data generation job.

The base source model for data generation jobs.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: AgentDataGenerationJobSource, FileDataGenerationJobSource, PromptDataGenerationJobSource, TracesDataGenerationJobSource

LLM model options for data generation jobs.

Token usage information for a data generation job.

Represents a reference to a blob for consumption.

Dataset output for a data generation job.

Dataset source for evaluator generation jobs — reference to a dataset.

Custom LROPoller for data generation job operations.

Reference to a versioned Foundry Dataset.

DatasetVersion Definition.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: FileDatasetVersion, FolderDatasetVersion

A deleted agent Object.

A deleted agent version Object.

Response for deleting a memory item from a memory store.

DeleteMemoryStoreResult.

A deleted skill.

A deleted skill version.

Model Deployment Definition.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ModelDeployment

A single dimension — one independent, measurable quality dimension within a rubric evaluator's scoring blueprint.

Identifiers returned after a routine dispatch is queued.

Embedding configuration class.

EmptyModelParam.

Endpoint-based evaluator definition. The customer owns and hosts an HTTP endpoint that implements the evaluation contract. The evaluator references a Project Connection by name; the connection stores the endpoint URL and credentials (API Key or Entra ID). At execution time, the service resolves the connection to obtain the endpoint URL and authentication details, then calls the endpoint for each evaluation row.

EntraAuthorizationScheme.

Entra ID credential definition.

Represents a CSV data source by file ID.

Represents a CSV data source for evaluation runs.

Result of the evaluation.

Metric comparison for a treatment against the baseline.

Comparison results for treatment runs against the baseline.

Summary statistics of a metric in an evaluation run.

Evaluation Comparison Request.

Insights from the evaluation comparison.

A sample from the evaluation result.

Evaluation rule model.

Evaluation action model.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ContinuousEvaluationRuleAction, HumanEvaluationPreviewRuleAction

Evaluation filter model.

Insights on set of Evaluation Results.

Insights from the evaluation run cluster analysis.

Evaluation task for the schedule.

Base class for targets with discriminator support.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: AzureAIAgentTarget, AzureAIModelTarget

Evaluation Taxonomy Definition.

Input configuration for the evaluation taxonomy.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: AgentTaxonomyInput

Request body for getting evaluator credentials.

Base evaluator configuration with discriminator.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: CodeBasedEvaluatorDefinition, EndpointBasedEvaluatorDefinition, PromptBasedEvaluatorDefinition, RubricBasedEvaluatorDefinition

Service-managed provenance artifacts produced by an evaluator generation job. Present only on EvaluatorVersion resources created via the generation pipeline. The combined-JSONL Foundry Dataset is read-only and resolves to a versioned dataset in a service-reserved namespace.

Caller-supplied inputs for an evaluator generation job.

Evaluator Generation Job resource — a long-running job that generates rubric-based evaluator definitions from source materials. On success, the result is the persisted EvaluatorVersion.

The base source model for evaluator generation jobs. Polymorphic over type.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: AgentEvaluatorGenerationJobSource, DatasetEvaluatorGenerationJobSource, PromptEvaluatorGenerationJobSource, TracesEvaluatorGenerationJobSource

Custom LROPoller for evaluator generation job operations.

Token consumption summary for an evaluator generation job. Populated when the job reaches a terminal state.

Evaluator Metric.

Evaluator Definition.

The external agent definition. Represents a third-party agent hosted outside Foundry (for example, on GCP or AWS). Registration is metadata-only: Foundry records the agent definition to light up observability experiences (traces, evaluations) over customer-emitted OpenTelemetry data.

The fabric data agent tool parameters.

A FabricIQ server-side tool.

A FabricIQ tool stored in a toolbox.

Field mapping configuration class.

Azure OpenAI file output for a data generation job.

File source for data generation jobs — Azure OpenAI file input.

FileDatasetVersion Definition.

File search.

A file search tool stored in a toolbox.

FixedRatioVersionSelectionRule.

FileDatasetVersion Definition.

A warning associated with a model.

Shell tool.

FunctionShellToolParamEnvironment.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ContainerAutoParam, FunctionShellToolParamEnvironmentContainerReferenceParam, FunctionShellToolParamEnvironmentLocalEnvironmentParam

FunctionShellToolParamEnvironmentContainerReferenceParam.

FunctionShellToolParamEnvironmentLocalEnvironmentParam.

Function.

FunctionToolParam.

A GitHub issue routine trigger.

Header-based secret authentication for a telemetry endpoint. The resolved secret value is injected as an HTTP header.

The hosted agent definition.

Hourly recurrence schedule.

Evaluation rule action for human evaluation.

HybridSearchOptions.

Image generation tool.

ImageGenToolInputImageMask.

Index resource Definition.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: AzureAISearchIndex, CosmosDBIndex, ManagedAzureAISearchIndex

InlineSkillParam.

Inline skill payload.

The response body for cluster insights.

A cluster of analysis samples.

Configuration of the model used in the insight generation.

The request of the insights report.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: AgentClusterInsightRequest, EvaluationComparisonInsightRequest, EvaluationRunClusterInsightRequest

The result of the insights.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: AgentClusterInsightResult, EvaluationComparisonInsightResult, EvaluationRunClusterInsightResult

A sample from the analysis.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: EvaluationResultSample

Insight task for the schedule.

Summary of the error cluster analysis.

Metadata about the insights.

Configuration specific to the invocations protocol.

Configuration specific to the WebSocket-based invocations protocol.

A manual payload used to test an invocations API routine dispatch.

Dispatches a routine through the raw invocations API. Exactly one of agent_name or agent_endpoint_id must be provided.

A manual payload used to test a responses API routine dispatch.

Dispatches a routine through the responses API. Exactly one of agent_name or agent_endpoint_id must be provided.

Local shell tool.

LocalSkillParam.

Adapter-specific metadata for LoRA models. Drives serving engine configuration at deployment time.

MCP tool.

MCP tool filter.

MCPToolRequireApproval.

An MCP tool stored in a toolbox.

ManagedAgentIdentityBlueprintReference.

Managed Azure AI Search Index Definition.

Configuration specific to the MCP protocol.

A single memory item stored in the memory store, containing content and metadata.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ChatSummaryMemoryItem, ProceduralMemoryItem, UserProfileMemoryItem

Represents a single memory operation (create, update, or delete) performed on a memory item.

A retrieved memory item from memory search.

Memory search options.

A tool for integrating memories into the agent.

Default memory store implementation.

Default memory store configurations.

Base definition for memory store configurations.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: MemoryStoreDefaultDefinition

Response for deleting memories from a scope.

A memory store that can store and retrieve user memories.

Usage statistics of a memory store operation.

Memory search response.

Memory update result.

Provides the status of a memory store update operation.

A set of delegated permission scopes requested from a single resource application.

Default and previously-published values used to pre-populate a Microsoft 365 publish request for a Foundry agent.

Response from publishing an agent to Microsoft 365 / Microsoft Teams.

The input definition information for a Microsoft Fabric tool as used to configure an agent.

Request to fetch credentials for a model asset.

Model Deployment Definition.

Sku information.

Represents a request for a pending upload of a model version.

Represents the response for a model pending upload request.

Represents a set of parameters used to control the sampling behavior of a language model during text generation.

Represents a set of parameters used to control the sampling behavior of a language model during text generation.

Source information for the model.

Model Version Definition.

Monthly recurrence schedule.

Namespace.

Credentials that do not require authentication.

One-time trigger.

Security details for OpenApi anonymous authentication.

authentication details for OpenApiFunctionDefinition.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: OpenApiAnonymousAuthDetails, OpenApiManagedAuthDetails, OpenApiProjectConnectionAuthDetails

The input definition information for an openapi function.

OpenApiFunctionDefinitionFunction.

Security details for OpenApi managed_identity authentication.

Security scheme for OpenApi managed_identity authentication.

Security details for OpenApi project connection authentication.

Security scheme for OpenApi managed_identity authentication.

The input definition information for an OpenAPI tool as used to configure an agent.

An OpenAPI tool stored in a toolbox.

Identifies the registered Foundry agent to optimize (request-only). Skills, tools, and system_prompt are specified in options.optimization_config.

An OTLP (OpenTelemetry Protocol) telemetry export endpoint.

Represents a request for a pending upload.

Represents the response for a pending upload request.

A memory item containing a procedure extracted from conversations.

ProgrammaticToolCallingParam.

Promotion metadata recorded when a candidate is deployed to a Foundry agent.

The prompt agent definition.

Configuration options for a text response from the model. Can be plain text or structured JSON data.

Prompt-based evaluator.

Prompt source for data generation jobs — inline text provided by the user.

Prompt source for evaluator generation jobs — inline text provided by the user.

Per-protocol configuration for the agent endpoint.

A record mapping for a single protocol and its version.

Configuration for Responsible AI (RAI) content filtering and safety features.

RankingOptions.

Reasoning.

Recurrence schedule model.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: DailyRecurrenceSchedule, HourlyRecurrenceSchedule, MonthlyRecurrenceSchedule, WeeklyRecurrenceSchedule

Recurrence based trigger.

Red team details.

RedTeamEvalRunDataSource.

Abstract class for target configuration.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: AzureOpenAIModelConfiguration

A reminder tool stored in a toolbox.

Represents the parameters for response retrieval item generation.

ResponseUsageInputTokensDetails.

ResponseUsageOutputTokensDetails.

Configuration specific to the responses protocol.

A routine definition returned by the service.

Base model for a routine action.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: InvokeAgentInvocationsApiRoutineAction, InvokeAgentResponsesApiRoutineAction

Optional authorization configuration for a routine dispatch.

Base model for a manual dispatch payload.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: InvokeAgentInvocationsApiDispatchPayload, InvokeAgentResponsesApiDispatchPayload

A single routine run returned from the run history API.

Base model for a routine trigger.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: CustomRoutineTrigger, GitHubIssueRoutineTrigger, ScheduleRoutineTrigger, TimerRoutineTrigger

Rubric-based evaluator definition — stores dimensions produced by the generate API. Used for both quality and safety evaluators.

A non-fatal advisory produced during rubric evaluator generation when resolved inputs are technically valid but likely too weak to produce a high-quality rubric. Read-only; service-generated. Persisted with the terminal EvaluatorGenerationJob.

Shared Access Signature (SAS) credential definition.

Schedule model.

A recurring cron-based routine trigger.

Schedule run model.

Schedule task model.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: EvaluationScheduleTask, InsightScheduleTask

Session defaults applied to sessions created for a hosted agent version.

A single entry in a directory listing.

Response from uploading a file to a session sandbox.

A single Server-Sent Event frame emitted by the hosted agent session log stream.

Each frame contains an event field identifying the event type and a data field carrying the payload as plain text. Although the current data payload is JSON-formatted, its schema is not contractual — additional keys may appear and the format may change over time. Clients should treat data as an opaque string and optionally attempt JSON parsing.

New event types may be added in the future. Clients should gracefully ignore unrecognized event types.

Wire format:

   event: log
   data: {"timestamp":"2026-03-10T09:33:17.121Z","stream":"stdout","message":"Starting server on port 18080"}

   event: log
   data: {"timestamp":"2026-03-10T09:34:52.714Z","stream":"status","message":"Successfully connected to container"}

The sharepoint grounding tool parameters.

The input definition information for a sharepoint tool as used to configure an agent.

A shell tool stored in a toolbox. This model is additive to toolbox configuration and does not modify the OpenAI tool contract or existing toolbox tool definitions.

The options for a data generation job with SimpleQnA type.

The options for a task generation data generation job. Use with multiturn evaluation scenarios and with prompt, file, or agent sources. Generated dataset rows include fields such as id, category, test_case_description, and desired_num_turns.

A skill resource.

Inline content for defining a simple skill without uploading files. Follows the agentskills.io SKILL.md specification.

SkillReferenceParam.

A specific version of a skill.

Specific apply patch tool choice.

Specific shell tool choice.

SpecificProgrammaticToolCallingParam.

An structured input that can participate in prompt template substitutions and tool argument binding.

A structured output that can be produced by the agent.

Represents a data source for target-based completion evaluation configuration.

Taxonomy category definition.

Taxonomy sub-category definition.

Customer-supplied telemetry configuration for exporting container logs, traces, and metrics.

A telemetry export endpoint configuration.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: OtlpTelemetryEndpoint

Authentication configuration for a telemetry endpoint.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: HeaderTelemetryEndpointAuth

AzureAIEvaluatorGrader.

An object specifying the format that the model must output. Configuring { "type": "json_schema" } enables Structured Outputs, which ensures the model will match your supplied JSON schema. Learn more in the Structured Outputs guide. The default format is { "type": "text" } with no additional options. Not recommended for gpt-4o and newer models:* Setting to { "type": "json_object" } enables the older JSON mode, which ensures the message the model generates is valid JSON. Using json_schema is preferred for models that support it.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: TextResponseFormatJsonObject, TextResponseFormatJsonSchema, TextResponseFormatText

JSON object.

JSON schema.

Text.

A one-shot timer routine trigger.

A tool that can be used to generate a response.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: A2ATool, A2APreviewTool, ApplyPatchToolParam, AzureAISearchTool, AzureFunctionTool, BingCustomSearchPreviewTool, BingGroundingTool, BrowserAutomationPreviewTool, CaptureStructuredOutputsTool, CodeInterpreterTool, ComputerTool, ComputerUsePreviewTool, CustomToolParam, MicrosoftFabricPreviewTool, FabricIQPreviewTool, FileSearchTool, FunctionTool, ImageGenTool, LocalShellToolParam, MCPTool, MemorySearchPreviewTool, NamespaceToolParam, OpenApiTool, ProgrammaticToolCallingParam, SharepointPreviewTool, FunctionShellToolParam, ToolSearchToolParam, WebIQPreviewTool, WebSearchTool, WebSearchPreviewTool, WorkIQPreviewTool

Allowed tools.

Indicates that the model should use a built-in tool to generate a response. Learn more about built-in tools.

Indicates that the model should use a built-in tool to generate a response. Learn more about built-in tools.

Indicates that the model should use a built-in tool to generate a response. Learn more about built-in tools.

Indicates that the model should use a built-in tool to generate a response. Learn more about built-in tools.

Custom tool.

Indicates that the model should use a built-in tool to generate a response. Learn more about built-in tools.

Function tool.

Indicates that the model should use a built-in tool to generate a response. Learn more about built-in tools.

MCP tool.

How the model should select which tool (or tools) to use when generating a response. See the tools parameter to see how to specify which tools the model can call.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ToolChoiceAllowed, SpecificApplyPatchParam, ToolChoiceCodeInterpreter, ToolChoiceComputer, ToolChoiceComputerUse, ToolChoiceComputerUsePreview, ToolChoiceCustom, ToolChoiceFileSearch, ToolChoiceFunction, ToolChoiceImageGeneration, ToolChoiceMCP, SpecificProgrammaticToolCallingParam, SpecificFunctionShellParam, ToolChoiceWebSearchPreview, ToolChoiceWebSearchPreview20250311

Indicates that the model should use a built-in tool to generate a response. Learn more about built-in tools.

Indicates that the model should use a built-in tool to generate a response. Learn more about built-in tools.

Per-tool configuration that controls tool visibility and search behavior.

Description of a tool that can be used by an agent.

Description of a tool that can be used by an agent.

A project connection resource.

Tool search tool.

A toolbox search tool stored in a toolbox.

The options for a data generation job with ToolUse type. Used only for fine-tuning scenarios.

A toolbox that stores reusable tool definitions for agents.

Policy configuration for a toolbox, including content safety and other governance settings.

A toolbox search tool stored in a toolbox.

An automatically provisioned container environment for a shell tool stored in a toolbox.

An existing container environment for a shell tool stored in a toolbox.

An execution environment for a shell tool stored in a toolbox. This environment model is scoped to toolbox configuration and does not modify the OpenAI shell environment contract.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ToolboxShellContainerAutoEnvironment, ToolboxShellContainerReferenceEnvironment

Network access policy for an automatically provisioned toolbox shell container.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ToolboxShellNetworkPolicyDisabled

A network policy that disables outbound access from a toolbox shell container.

A skill source included in a toolbox.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: ToolboxSkillReference

A reference to an existing skill to include in a toolbox.

An abstract representation of a tool stored in a toolbox.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: A2AToolboxTool, A2APreviewToolboxTool, AzureAISearchToolboxTool, BrowserAutomationPreviewToolboxTool, CodeInterpreterToolboxTool, FabricIQPreviewToolboxTool, FileSearchToolboxTool, MCPToolboxTool, OpenApiToolboxTool, ReminderPreviewToolboxTool, ShellToolboxTool, ToolSearchToolboxTool, ToolboxSearchPreviewToolboxTool, WebIQPreviewToolboxTool, WebSearchToolboxTool, WorkIQPreviewToolboxTool

A specific version of a toolbox.

The options for a data generation job with Traces type.

Traces source for data generation jobs — conversation traces from Application Insights.

Traces source for evaluator generation jobs — conversation traces from Application Insights.

Represents a data source for evaluation runs that operate over Agent traces stored in Application Insights.

Base model for Trigger of the schedule.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: CronTrigger, OneTimeTrigger, RecurrenceTrigger

Custom LROPoller for Memory Store update operations.

Request body for updating a model version. Only description and tags can be modified.

UpdateToolboxRequest.

A memory item specifically containing user profile information extracted from conversations, such as preferences, interests, and personal details.

Version indicator determining which agent version backs the session.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: VersionRefIndicator

Version indicator that references a specific agent version by name.

VersionSelectionRule.

You probably want to use the sub-classes and not this class directly. Known sub-classes are: FixedRatioVersionSelectionRule

VersionSelector.

A WebIQ server-side tool.

A WebIQ tool stored in a toolbox.

Web search approximate location.

A web search configuration for bing custom search.

Web search preview.

Web search.

WebSearchToolFilters.

A web search tool stored in a toolbox.

Weekly recurrence schedule.

A WorkIQ server-side tool.

A WorkIQ tool stored in a toolbox.

The workflow agent definition. Microsoft Foundry is retiring workflows on December 1, 2026. If you're looking to build new workflows, use Microsoft Agent Framework. To migrate existing workflows, see the Migration guide.

Supported A2A protocol versions.

An access boundary for the activity protocol.

Type of AgentBlueprintReferenceType.

Type of AgentEndpointAuthorizationSchemeType.

Type of AgentEndpointProtocol.

The status of an agent identity, applicable to both the agent instance identity and the agent blueprint.

Identifies where an Agent Insights overview came from.

The Prompt surface changed by a proposed fix.

The customer-renderable kind of an agent insight's proposed fix.

The trigger that started an agent insight run.

The severity of an agent insight.

The lifecycle status of an agent insight.

Type of AgentKind.

Type of AgentObjectType.

Discriminator values for the dataset input union.

The status of an agent session.

The operational state of an agent.

Indicates the source of an agent's operational state. Empty when the state is not derived from a specific source.

The provisioning status of an agent version.

Strategies for attacks.

Available query types for Azure AI Search tool.

Type of CallableToolAllowedCaller.

How package dependencies are resolved at deployment time for a code-based hosted agent.

Type of ComputerEnvironment.

The Type (or category) of the connection.

Type of ContainerMemoryLimit.

Type of ContainerNetworkPolicyParamType.

Type of ContainerSkillType.

The credential type used by the connection.

Type of CustomToolParamFormatType.

The supported output file types for a data generation job.

The supported scenarios for a data generation job.

The supported source types for data generation jobs.

The supported data generation job types.

Enum to determine the type of data.

Days of the week for recurrence schedule.

Type of DeploymentType.

The type of digital worker.

The level at which evaluation is performed.

Type of the evaluation action.

Type of the evaluation rule event.

Type of the evaluation taxonomy input.

The category of the evaluator.

The type of evaluator definition.

The supported source types for evaluator generation jobs.

The direction of the metric indicating whether a higher value is better, a lower value is better, or neutral.

The type of the evaluator.

The type of the evaluator.

The artifact profile category.

Signals detected in the model artifact.

The source type of the model.

Warning code for model artifacts.

The weight type of the model.

Type of FunctionShellToolParamEnvironmentType.

Category of a warning surfaced on a generated evaluator version. Extensible so new warning categories (e.g., safety, output quality) can be introduced without a breaking change.

Known GitHub issue events that can fire a routine.

Type of GrammarSyntax1.

Type of ImageGenAction.

Type of IndexType.

Control how much effort the model will exert to match the style and features, especially facial features, of input images. This parameter is only supported for gpt-image-1 and gpt-image-1.5 and later models, unsupported for gpt-image-1-mini. Supports high and low. Defaults to low.

The request of the insights.

Extensible status values shared by Foundry jobs.

Memory item kind.

Memory operation kind.

The type of memory store implementation to use.

Type of MemoryStoreObjectType.

Status of a memory store update operation.

The publish scope for the generated Microsoft Teams app.

Authentication type for OpenApi endpoint. Allowed types are:

• Anonymous (no authentication required)

• Project Connection (requires project_connection_id to endpoint, as setup in AI Foundry)

• Managed_Identity (requires audience for identity based auth).

Enum describing allowed operation states.

Type of PageOrder.

The type of pending upload.

The Microsoft Agent Certification review status of the Microsoft 365 store title published for an agent.

Type of RankerVersionType.

Constrains effort on reasoning for reasoning models. Currently supported values are none, minimal, low, medium, high, xhigh, and max. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response. Not all reasoning models support every value. See the reasoning guide for model-specific support.

Type of ReasoningModeEnum.

Recurrence type.

Risk category for the attack objective.

The discriminator values supported for routine actions.

Known source paths that can produce a routine run.

The supported identities for routine dispatch authorization.

The discriminator values supported for manual routine dispatch payloads.

Known lifecycle phases recorded for a routine run.

The discriminator values supported for routine triggers.

Stable searchable machine-readable warning code for a rubric-generation input-quality warning. Values are snake_case; clients must tolerate additional service-defined identifiers.

Advisory severity for a rubric-generation input-quality warning. Initial value set: warning.

Warning source attribution for a rubric-generation input-quality warning. Per-source values (prompt, agent, dataset) match the source category visible to the generation runtime. aggregate is a synthetic value used only for warnings computed across successfully resolved sources. traces is not exposed because trace sources resolve into dataset content upstream.

The type of sample used in the analysis.

Schedule provisioning status.

Type of the task.

Type of SearchContentType.

Type of SearchContextSize.

Known SSE event types emitted by the hosted agent session log stream. Additional event types may be introduced in future versions.

The supported question types for SimpleQnA data generation jobs used for fine-tuning scenarios.

The type of telemetry data to export.

The type of authentication for a telemetry endpoint.

The kind of telemetry export endpoint.

The transport protocol for telemetry export.

Type of TextResponseFormatConfigurationType.

Type of ToolChoiceParamType.

Type of ToolSearchExecutionType.

Type of ToolType.

Supported tool types for tools stored in a toolbox.

Treatment Effect Type.

Type of the trigger.

The type of version indicator used to determine the agent version backing a session.

Type of VersionSelectorType.