Skip to main content Crate rust_ml_runtime Copy item path Source pub use ml_runtime_backend ;pub use ml_runtime_common ;pub use ml_runtime_inference ;pub use ml_runtime_model ;pub use ml_runtime_provider ;ApiKey A bearer credential. Debug is redacted and the type is not serializable. BackendRegistry BuiltinCapabilityProvider BuiltinModelRegistry CancellationToken A token which can be used to signal a cancellation request to one or more
tasks. Capability CapabilityAvailability CapabilityExecutionMetadata CapabilityLimits CapabilityPackageManifest CapabilityPolicy CapabilityRegistry CapabilityResolution ChatMessage ChatMetadata Execution evidence recorded by the runtime/provider (not by the model). ChatRequest ChatResponse ChatStreamSummary Terminal stream summary. It deliberately omits the full text, so a stream
never forces the runtime to buffer unbounded output. CompiledArtifactManifest DecisionDependency DecisionExecution DecisionExecutionDiagnostics DecisionExecutionPlan DecisionExecutionPolicy Application constraints. These bound execution without prescribing a backend strategy. DecisionGraphRequest DecisionModelCapabilities Backend/model execution facts used by the structured-decision planner. DecisionNode One decision node and the predicates that control whether it is required. DecisionOption DecisionPlanStage DecisionPlanner DecisionProvenance DecisionQuestion DecisionRequest A model-neutral request for one or more typed decisions about immutable input state. DecisionResult ExecutionMetadata ExecutionRecord ExecutionRequest ExecutionResult FilesystemModelCatalog GenerationCapabilities Optional features of a generation provider. The OpenAI-compatible adapter
cannot discover these, so they are configured and enforced up front. GenerationModelInfo GenerationOptions InferenceOptions InferenceRequest InferenceResult InstallationManifest InstalledModelRecord LocalProviderPackage ManifestCapabilityProvider ModelCapability ModelDescription ModelDescriptor Catalog discovery information. A descriptor is metadata, not a loaded resource. ModelDiagnostic ModelDoctorReport ModelId A validated, stable model name and version used by catalogs and installers. ModelIdentity ModelLifecycleResult ModelRegistry ModelResource ModelSpec An explicit backend-independent model definition and artifact location. OpenAiCompatibleConfig OpenAiCompatibleProvider PackageCapability PackageManifest PlannedDecisionResult Result returned by the capability-aware executor. Flattening preserves the
existing JSON shape while keeping DecisionResult source-compatible. ProviderDescriptor ProviderManifest ProviderRegistry RegisteredModelInstaller ResolutionCandidate ResolutionConstraints Runtime Coordinates catalog resolution, routing, model lifecycle, and inference. RuntimeBuilder Explicit construction and extension surface for an application-owned Runtime . RuntimeCapabilities RuntimeConfig RuntimeEnvironment RuntimeMetrics RuntimeMetricsSnapshot RuntimeModelHandle Opaque runtime-owned reference to a loaded model resource. RuntimeSelection RuntimeSnapshot RuntimeStatus Tensor ToolCall A tool-call proposal produced by a model. It is not evidence that the
operation ran, was authorized, or would succeed; the calling application
decides whether and how to execute it. ToolDefinition A tool the model may propose to call. The runtime never executes it: the
schema constrains the proposed arguments and grants no authority. TypedDecision Usage Token accounting as reported by the provider; never estimated. AvailabilityState ChatRole ChatStreamEvent CompiledStatus DecisionDependencyCondition DecisionExecutionStrategy DecisionType DecisionValue DeviceKind ExecutionCompletion ExecutionLocation ExecutionPolicy Determines where the runtime attempts to execute a request and whether it may fall back. FinishReason Why generation stopped, when the provider says. Other preserves an
unrecognized provider value verbatim. InferenceStreamEvent Input InstalledModelStatus ModelArchitecture ModelFormat ModelLifecycleState ModelLocation ModelReference A request-time model reference: either a catalog identity or an explicit specification. Output Platform ResponseFormat Requested output shape. JsonSchema is a request to the provider; it is
never a guarantee. The runtime validates the result separately. RuntimeError Stable, structured failures returned by the public runtime API. StructuredOutput Result of checking generated text against the requested format. ToolChoice Provider-neutral tool selection policy. Absent means the provider default. VERSION CapabilityAuthorizer CapabilityProvider DecisionModel Loaded model-neutral typed-decision model. DecisionModelProvider Backend extension capable of loading a local typed-decision artifact. GenerationProvider A provider of generative inference. Separate from crate::Provider ,
which serves typed inference. default_model_root ExecutionId InferenceStream Bounded asynchronous stream of normalized runtime-owned inference events. RequestId RuntimeResult