UnrealMCP — AI Bridge for Unreal Engine 5

List materials that can be applied to objects. Args: search_path: Content Browser root to scan (e.g. "/Game/Weapons/") include_engine_materials: Include /Engine/ materials in results filter: Case-insensitive substring match on material asset name max_results: Cap on returned entries. 0 = unlimited (use sparingly)

Get all default property values from a Blueprint's generated class CDO. Unlike get_blueprint_variable_details (which only shows BP-defined variables), this reads the CDO and returns ALL editable properties — including C++ parent ones. Args: filter: Substring to filter property names (case-insensitive) include_inherited: Include properties from C++ parent (default True)

Get detailed info about Blueprint functions (empty name = all functions).

Get material slot info for a Blueprint's mesh components.

Get detailed info about Blueprint variables (empty name = all variables).

Read ALL property values from a data asset (no visibility filter). Args: filter: Substring to filter property names (case-insensitive) include_inherited: Include properties from parent C++ classes (default True)

Get a single row from a Data Table by its row name.

Get all rows and their field data from a Data Table asset.

Get column names and types defined by the Data Table's row struct.

Look up pin names and editable properties for a node type WITHOUT creating one. Returns input pins, output pins, and all editable properties with types and defaults. Use this BEFORE creating nodes to know exact pin names and avoid connection errors. Special node types: - Custom: Returns additional custom_hlsl_schema with code/inputs/outputs documentation. - MaterialFunctionCall: Pass function_path to see the actual inputs/outputs of that function. Without it, pins will be empty since they depend on which function is loaded. Use search_material_functions to find functions first. - Substrate nodes (SubstrateSlabBSDF, SubstrateShadingModels, etc.): Returns substrate_note explaining how to connect to FrontMaterial pin. Args: type_name: Short type name (e.g. "Multiply", "TextureSample", "Custom", "MaterialFunctionCall", "SubstrateSlabBSDF", "SubstrateShadingModels") function_path: For MaterialFunctionCall only — path to the material function to load (e.g. "/Engine/Functions/Engine_MaterialFunctions03/Blends/Blend_Overlay"). Populates the node with the function's actual input/output pins.

Read all properties, triggers, and modifiers of an Input Action. Returns value_type, consume_input, trigger_when_paused, accumulation_behavior, and full details of all triggers and modifiers with their properties.

Read all key mappings from an Input Mapping Context. Returns each mapping's key, action, triggers, and modifiers with full property details.

Read all traits from a Mass Entity Config asset with their properties expanded. Unlike get_data_asset_properties (which shows trait object paths), this tool expands each trait's editable properties with current values. Args: asset_path: Path to MassEntityConfigAsset (e.g. "/Game/.../DA_ConveyorMassConfig_Mk1") Returns per trait: index, name, class, class_path (for _ClassName), and all editable properties with current values.

Get shader compilation errors for a material. Returns error messages and the node indices that caused them. Recompiles first by default to get fresh errors.

Get detailed info for a single node, including all available input/output pins. Use this before connecting nodes to discover exact pin names.

Inspect a Material Function: inputs, outputs, internal nodes. Returns all FunctionInput pins (name, type, default value, sort priority), all FunctionOutput pins (name, connections), and a summary of internal nodes.

Read expression nodes in a material graph. Use verbosity to control response size: - "summary": index, type, position only (~30 tokens/node) - "connections": + input connections (default, good balance) - "full": + properties + available pins (large response) Use type_filter to only return nodes matching a type substring (e.g. "Parameter" returns only parameter nodes).

Inspect a material's properties. Always returns: blend mode, shading model, two-sided, expression count. Use include (comma-separated) to request additional sections: - "parameters": scalar/vector/texture parameter names - "textures": used texture paths - "statistics": shader instruction counts Omit include to get only basic properties (minimal tokens). Pass include="parameters,textures,statistics" for everything.

Get all overridable parameters from a Material Instance (scalar, vector, texture, switch).

Query which expression node feeds each material output slot. Only lists slots that have something connected.

Get accumulated per-command token usage statistics. Returns commands sorted by max_estimated_tokens (worst offenders first). Each entry shows: call_count, avg/max/total estimated tokens, response chars. Enable debug mode first with set_mcp_debug(enabled=True).

Get all Niagara actors in the current level. Args: system_filter: Optional filter on the Niagara System name or path

Return the full editable-property schema of a UNiagaraDataInterface subclass. Walks the CDO via the modular FProperty introspector. Covers every DI: Array{Float,Float2,Float3,Float4,Color,Quat,Position,Matrix}, SkeletalMesh, StaticMesh, Spline, Curve, RenderTarget2D, VolumeTexture, CameraQuery, NeighborGrid3D, Grid2D, Texture, etc. Each field returns: name, kind, sub-type, display_name, tooltip, category, clamp/UI min/max where present, recursive nested struct schema, enum entries for enum-typed fields, etc. Use to discover what fields a DI supports before configuring or instantiating one. Args: class_: Class name with or without the "UNiagaraDataInterface" prefix. Examples: "ArrayFloat", "UNiagaraDataInterfaceSkeletalMesh", "Spline", "RenderTarget2D".

Get attributes defined on an emitter at a given scope. Returns name, type, scope, and source per attribute. Includes both rapid iteration parameters and well-known particle attributes (Position, Velocity, Color, etc.). Use filter to narrow by attribute name substring. Args: system_path: Path to the Niagara System asset emitter_name: Name of the target emitter scope: Attribute scope — "particle", "emitter", "system", "all" filter: Optional name filter (e.g. "Color", "Position", "Velocity", "Size")

Get emitters in a Niagara System. Returns name, unique_name, index, enabled state, sim_target, local_space, determinism, and renderer_count per emitter. Use filter to get specific emitters by name substring. Args: system_path: Path to the Niagara System asset filter: Optional name filter (e.g. "Spark", "Flame", "Trail")

Get all event handlers configured on an emitter. Args: system_path: Path to the Niagara System asset emitter_name: Name of the target emitter

Introspect every node inside a Niagara graph. Three resolver modes: 1. system_path + module_name — scratch pad module graph 2. system_path + emitter_name + script_usage — EMITTER STACK GRAPH (emitter_spawn|emitter_update|particle_spawn|particle_update), where stack modules appear as NiagaraNodeFunctionCall nodes 3. script_path — standalone UNiagaraScript asset Args: system_path: Niagara System asset path module_name: Scratch pad module name (mode 1) emitter_name + script_usage: Emitter stack graph (mode 2) script_path: Standalone script asset (mode 3) verbosity: "summary" | "connections" (default) | "full" type_filter: Case-insensitive substring on short class name name_filter: Case-insensitive substring on node title

Resolve the actual binding of each module input — Default/Local/Linked/Dynamic/Data. This answers "what is Spawn Count actually driven by?" — something get_niagara_module_inputs + get_niagara_rapid_iteration_parameters can't do on their own. Returns per-input: - mode: default | local | linked | dynamic | data | function_call | expression | unknown - value: literal string when mode=local - linked_parameter: full parameter name when mode=linked/data - script_path + function_name: when mode=dynamic/function_call - children: recursive — for dynamic inputs, lists their own bound inputs Uses FNiagaraStackGraphUtilities::GetStackFunctionInputs (exported) and replicates GetStackFunctionInputOverridePin manually (non-exported). Args: system_path: Niagara System asset emitter_name: Emitter in the system module_name: Module display name (e.g. "SpawnPerFrame") script_usage: emitter_spawn|emitter_update|particle_spawn|particle_update| system_spawn|system_update input_filter: Case-insensitive substring on input name max_depth: Recursive descent cap for dynamic-input children (default 3)

Get all input parameters for a specific module. Args: system_path: Path to the Niagara System asset emitter_name: Name of the target emitter module_name: Name of the module script_usage: Stack the module is in input_filter: Optional substring filter on input parameter names include_schema: If true, include massive structural type schemas for inputs

Check for outdated modules in an emitter and their available versions. Identifies modules that have newer versions available — a common source of Niagara warnings and compatibility issues. Use to find modules that need upgrading. Args: system_path: Path to the Niagara System asset emitter_name: Name of the target emitter filter: Optional module name filter (e.g. "Spawn", "Color", "Force")

Get modules in an emitter's script usage stack. Returns name, display_name, index, script_usage, script_path, and optionally input parameters per module. Use filter to get specific modules by function name. Args: system_path: Path to the Niagara System asset emitter_name: Name of the target emitter script_usage: Which stack — EmitterSpawn, EmitterUpdate, ParticleSpawn, ParticleUpdate, SystemSpawn, SystemUpdate, or "all" for every stack include_inputs: If true, include each module's input parameters filter: Optional function name filter (e.g. "Spawn", "Color", "Gravity", "Velocity")

Deep inspect a single Niagara node: full pin layout, connections, type-specific detail. Provide exactly one of node_index / node_class / node_id. Args: system_path + module_name: Resolve scratch pad module (one option) script_path: Resolve standalone script asset (alternative) node_index: Ordinal index into Graph->Nodes (stable per session) node_class: Short ("MapGet") or full ("NiagaraNodeParameterMapGet") node_id: FGuid string matching UEdGraphNode::NodeGuid

Get pin/property schema for a Niagara node type or script asset. For a node class (e.g. "CustomHlsl", "ParameterMapGet"), returns the editable UPROPERTY list via reflection. For CustomHlsl specifically, returns the full HLSL schema (like Material's get_expression_type_info does for the Custom node) with an example payload. For a script_path, loads the asset and returns its input/output parameters. Args: type: Node short name (e.g. "CustomHlsl", "ParameterMapGet") or full class name script_path: Optional script asset path — when set, returns script input/output info instead

Get live particle counts and emitter execution state from running preview. Shows how many particles are alive per emitter, total spawned count, execution state (Active/Inactive/Complete), and bounds info. Requires the system to be previewing in the Niagara editor or spawned in level. Args: system_path: Path to the Niagara System asset emitter_name: Optional — filter to a specific emitter's stats

Get the current timeline playback range and frame rate settings. Returns the start/end time in seconds and the frame rate configuration. Use to check if the playback range is too short (common cause of effects appearing to not play). Args: system_path: Path to the Niagara System asset

Get rapid iteration parameters (the actual configurable values on modules). These are the real module inputs — spawn rate, gravity, colors, sizes, etc. Returns parameter names, types, current values, and which module they belong to. Args: system_path: Path to the Niagara System asset emitter_name: Name of the target emitter script_usage: Stack filter — emitter_spawn, emitter_update, particle_spawn, particle_update, or all filter: Optional substring filter on parameter name (e.g. "Color", "SpawnRate", "Gravity")

Get detailed information about a renderer on an emitter. Args: system_path: Path to the Niagara System asset emitter_name: Name of the target emitter renderer_index: Index of the renderer (default 0, the first renderer)

Get editable properties of a renderer with current values and valid enum values. Token-efficient: use filter to get only specific properties instead of all. Property names can be used directly with set_niagara_renderer_property. Common filters: "Facing", "Material", "Sort", "Shadow", "Width", "Shape", "UV" Args: system_path: Path to the Niagara System asset emitter_name: Name of the target emitter renderer_index: Renderer index (default 0) filter: Substring filter on property name — strongly recommended to save tokens

Return the full graph context-menu actions for a scratch pad module. Wraps UEdGraphSchema_Niagara::GetGraphActions — the same call the editor uses to populate the right-click "add node" menu for a Niagara script graph. Each entry has display_name, category, tooltip, keywords, and template_class. Args: system_path: Niagara System asset path module_name: Scratch pad module name (the actions are specific to its graph) filter: Optional substring applied to display_name/category/tooltip/keywords max_results: Cap on returned entries

Read the details-panel properties of a Niagara script. Returns Category, Description, Keywords, ModuleUsageBitmask, ProvidedDependencies, RequiredDependencies, LibraryVisibility, bDeprecated, DeprecationMessage, bExperimental, ExperimentalMessage, NumericOutputTypeSelectionMode, ScriptMetaData, ConversionUtility. All properties live on FVersionedNiagaraScriptData; output shape matches NiagaraIntrospection::SerializeProperty (kind + value + schema fields).

Get compilation errors, warnings, and validation issues for a Niagara System. Returns issues found during compilation and validation. Use to diagnose why a system isn't working, find outdated modules, or check for configuration problems. Args: system_path: Path to the Niagara System asset emitter_name: Optional — filter to issues from a specific emitter severity: Filter by severity — "error", "warning", "info", or "all" (default)

Get detailed information about a Niagara System. Returns system name, emitter list, user parameters, and compilation status. Use filter to narrow emitters and parameters by name substring. Args: system_path: Path to the Niagara System asset include: Comma-separated sections — emitters, parameters, compilation, all filter: Optional name substring filter on emitters and parameters (e.g. "Spark", "Color", "Flame")

Get user-exposed parameters from a Niagara System asset or a level actor's component. Returns name, type, and value_type per parameter. Provide either system_path (asset) or actor_name (level instance). Use filter to narrow by parameter name substring. Args: system_path: Path to the Niagara System asset actor_name: Name of an actor in the level with a NiagaraComponent filter: Optional name filter (e.g. "Color", "Intensity", "Size")

Return valid classes/structs/enum-values for a property dropdown. Mirrors the exact enumeration the Unreal Details panel uses: - Instanced UObject* / TArray<UObject*> -> GetDerivedClasses(PropertyClass) - TSubclassOf<T> -> GetDerivedClasses(MetaClass) - TArray<FInstancedStruct> -> UScriptStruct descendants of BaseStruct - Enum properties -> all named entries + values Supports dot-notation: property_path="Config.Traits" navigates into nested structs. IMPORTANT: Use 'filter' to avoid huge responses — without it, properties like Fragments return 200+ entries (~10k tokens). Args: class_name: Class or struct name (e.g. "MassAssortedFragmentsTrait") property_path: Dot-separated path (e.g. "Config.Traits", "Fragments") filter: Case-insensitive substring filter on name and path include_abstract: Include abstract base classes (default False)

Get all currently selected actors in the editor viewport.

Get currently selected assets in the Content Browser.

Read layout slot properties (type depends on parent: CanvasPanelSlot, HorizontalBoxSlot, etc.).

Get property bindings with full path details. Each binding shows source/target node names, struct IDs, and property paths (e.g. "TargetLocation.X"). Optionally filter to bindings involving a specific node. Args: asset_path: Content path to the StateTree asset node_guid: Only show bindings involving this node (optional)

Get all global evaluators with their properties and instance data. Args: asset_path: Content path to the StateTree asset filter: Optional substring filter on evaluator class name

Get complete StateTree data in one call with verbosity control. Three verbosity levels: - "summary": Counts only (state_count, evaluator_count, etc.) — minimal tokens - "standard": All states (recursively detailed), evaluators, global tasks, parameters - "full": Everything from standard + property bindings inlined on every node Use 'sections' to request only specific parts and save context: - "states" — state hierarchy with tasks, conditions, transitions - "evaluators" — global evaluators - "global_tasks" — global tasks - "parameters" — tree-level parameters - "bindings" — all property bindings as a flat list Comma-separate multiple sections: "states,bindings" Args: asset_path: Content path to the StateTree asset verbosity: "summary", "standard", or "full" sections: Comma-separated section filter (empty = all sections)