UnrealMCP — AI Bridge for Unreal Engine 5

Get all global tasks with their properties. Args: asset_path: Content path to the StateTree asset filter: Optional substring filter on task class name

Get compact overview of a StateTree: schema, state count, evaluator count, etc. Args: asset_path: Content path to the StateTree asset

Get detailed info for a single node (task/evaluator/condition) by GUID. Returns the node's class, all properties, instance data, and where it lives in the tree (which state, evaluator list, or global tasks). Args: asset_path: Content path to the StateTree asset node_guid: GUID of the node (from get_statetree_state or get_statetree_evaluators)

Get tree-level parameters (name, type, value). Args: asset_path: Content path to the StateTree asset

Get detailed info for a single state: properties, tasks, conditions, transitions. Identify by index path (e.g. "0/1/2"), name, or GUID. Args: asset_path: Content path to the StateTree asset state: Index path (e.g. "0" for first root, "0/1" for second child of first root) state_name: State name (case-insensitive) state_guid: State GUID string

Get state hierarchy as a JSON tree. Each state shows name, type, ID, task/transition counts. Use max_depth to limit recursion (0=roots only, 1=roots+children, -1=unlimited). Use name_filter for case-insensitive substring match on state names. Args: asset_path: Content path to the StateTree asset max_depth: Maximum depth to recurse (-1 for unlimited) name_filter: Only include states whose name contains this substring

Get all valid transition targets for a StateTree. Returns two lists: - 'states': All states with name, GUID, index_path, and type (for GotoState transitions) - 'meta_targets': Built-in targets (None, TreeSucceeded, TreeFailed, TreeStopped, NextSibling, NextState, NextSelectableState) Use this before add_statetree_transition or set_statetree_transition_property to discover valid target_state values. Args: asset_path: Content path to the StateTree asset

Read all property values from a widget (supports all FProperty types).

Read the complete widget hierarchy from a Widget Blueprint's Designer. Returns a recursive JSON tree: name, class, is_panel, is_variable, slot_class, and children[] for panels.

Get current editor level info (world name, actor count, actor list).

Check if the UE5 editor bridge is running and responsive.

Import an external file (texture, mesh, FBX, etc.) into the Content Browser. Args: source_file: Absolute path to file on disk (e.g. "C:/textures/wood.png") destination_path: Content Browser path (e.g. "/Game/Textures") destination_name: Custom asset name (default: source filename without extension) replace_existing: Overwrite if asset already exists (default True)

Import multiple files into the Content Browser in one batch. Provide EITHER 'files' (explicit list) OR 'source_directory' (scan a folder). Args: destination_path: Content Browser destination (e.g. "/Game/Textures") files: List of absolute file paths to import source_directory: Absolute path to a folder — all matching files will be imported extensions: File extensions to include when scanning a directory (e.g. ["png", "jpg", "fbx"]). If omitted with source_directory, all files are imported. replace_existing: Overwrite existing assets (default True)

Auto-layout all nodes using Unreal's built-in arrangement algorithm.

Link a module input to a Niagara parameter (user, system, emitter, or particle scope). Args: system_path: Path to the Niagara System asset emitter_name: Name of the target emitter module_name: Name of the module input_name: Name of the input to link script_usage: Stack the module is in linked_parameter: Full parameter name including namespace (e.g. "User.MyParam", "Emitter.Age", "Particles.Velocity")

List assets in a Content Browser directory. Args: path: Content path (e.g. "/Game", "/Game/Materials") class_filter: Optional class shortcut (e.g. "material", "blueprint")

List all UDataAsset subclasses currently loaded in the editor. Returns name, full class path, parent class, whether it is a UPrimaryDataAsset, and the number of editable properties.

List data assets in a content path, optionally filtered by class. Args: path: Content Browser path (e.g. "/Game/Data") class_filter: Class name filter (e.g. "UItemData"). Matches any UDataAsset if empty.

List all UInputAction assets, optionally filtered by name and path. Returns name, path, value_type, trigger_count, modifier_count for each.

List available FKey values (keyboard, mouse, gamepad keys). Args: filter: Substring filter on key name or display name category: Filter by category — keyboard | mouse | gamepad | analog | digital max_results: Max keys to return (default 500)

List all UInputMappingContext assets, optionally filtered. Returns name, path, mapping_count, description for each.

List available material expression node types. Use filter to narrow results (e.g. "texture", "parameter", "math", "noise"). Set max_results to limit output (0 = unlimited). Set include_details=false for compact output (type names only).

List all available UInputModifier subclasses with their editable properties. Useful for discovering valid modifier_type values. Returns name, display_name, parent class, and editable property names/types.

List parameters that can be bound to a Map Get / Map Set pin. Returns well-known Engine.* parameters and Particles.* attributes unconditionally. When system_path + module_name are supplied, also adds: * User.* parameters from the system's exposed parameter collection * Rapid-iteration parameters from every emitter script * The target scratch pad module's own script graph variables (Module.* / Input.* / Transient.*) Args: filter: Optional case-insensitive substring filter on parameter name namespace: "all" (default), "engine", "particles", "user", "module", "input", "transient", "rapid_iteration" max_results: Cap on returned entries (default 500) system_path: Optional system path to pull per-emitter + user params module_name: Optional scratch pad module to pull script-local params

Enumerate member functions on a Niagara data interface class. These are the functions you see in the right-click "Functions" submenu on a DI pin in the editor — built-in DI members like Array.Length, Array.Get, Array.Add, etc. They aren't UNiagaraScript assets so they don't show up in search_niagara_functions. Pass the result as function_name to add_niagara_graph_node(node_type="DataInterfaceFunction"). Each entry reports: - name: pass to function_name param - inputs / outputs: pin schema with names + Niagara types - supports_cpu / supports_gpu / read_function / write_function - description (when DI provides it) Authoritative source: UNiagaraDataInterface::GetFunctionSignatures (NIAGARA_API exported, NiagaraDataInterface.h:681). Args: di_class: Short ("NiagaraDataInterfaceArrayPosition") or full path filter: Optional substring on function name (e.g. "Length", "Get") include_pins: Default True — include input/output pin schema

List available Niagara Data Interfaces (Grid2D, Grid3D, AudioSpectrum, etc.). Args: filter: Optional substring filter on data interface name

List available emitter templates for use with add_niagara_emitter. Args: category: Filter by category — "all", "sprite", "mesh", "ribbon", "light", "audio"

Reproduce the stack-UI source dropdown for a specific input. Returns the exact options the editor offers when you click the dropdown on a stack input (screenshot reference: "Change Source" menu with Dynamic Inputs / Link Inputs / Make sections). Output: - input_type: the Niagara type of the target input (so you can filter) - dynamic_inputs: every UNiagaraScript with Usage=DynamicInput, with script_path ready to pass to set_niagara_dynamic_input - link_parameters: well-known engine/particle parameters + this system's user parameters, grouped by namespace Discovery is via: - FNiagaraEditorUtilities::GetFilteredScriptAssets (exported) for dynamic inputs - FNiagaraConstants::GetEngineConstants / GetCommonParticleAttributes for engine-side parameters - UNiagaraSystem::GetExposedParameters for the system's own user params

List available Niagara module scripts for use with add_niagara_module. Args: category: Filter by category — "all", "spawn", "update", "forces", "rendering", "collision", "audio", "events" search: Optional substring search on module name max_results: Maximum number of results to return

Enumerate every Niagara node type that can be added to a graph. Combines three sources, mirroring what the right-click menu shows in the editor: * Node classes (UNiagaraNode subclasses) — Map Get, Map Set, Custom HLSL, If, etc. * Script assets — module / dynamic_input / function scripts from the Asset Registry * Data Interface classes (UNiagaraDataInterface subclasses) Filter-first to keep output compact. Each entry includes kind, display_name, category, tooltip/keywords, and a class_path or script_path for use with add_niagara_module / add_niagara_node_pin. Args: filter: Case-insensitive substring applied to name/display/category/keywords/tooltip kind: "all" (default), "node_class", "module_script", "dynamic_input_script", "function_script", "data_interface" include_engine: When false, script asset scan is restricted to /Game and /Niagara max_results: Cap on total entries returned (default 500)

Enumerate all valid UNiagaraNodeOp operations from FNiagaraOpInfo registry. This is the authoritative source the editor's context menu reads from — use it BEFORE calling add_niagara_graph_node with node_type="Op" so you pass a real op_name (e.g. "Numeric::Mul" vs the invalid "Numeric::Multiply"). Each entry reports: name (for op_name param), friendly_name, category, description, keywords, alternate_name, compact_name, supports_added_inputs, and — when include_pins=True — full input/output pin schema with types and defaults. Args: filter: Case-insensitive substring, matches name / friendly_name / category / description / keywords / alternate name. E.g. "multiply" finds "Numeric::Mul" via its friendly name. category: Exact category match (e.g. "Math", "Comparison", "Boolean", "Vector", "Quaternion"). Use the all_categories field in a first response to discover valid values. exact_name: Return the single op whose internal Name equals this string (bypasses other filters). include_pins: Include inputs/outputs per op (default True). max_results: Cap on returned entries (default 500).

Live query of FNiagaraTypeRegistry — every type the editor's User Parameters "+" menu can add, including plugin-registered Data Interfaces. Args: scope: Which registry slice to query. "user" (default) = types addable as User.* parameters. Other scopes: "system", "emitter", "particle", "parameter", "payload", "numeric", "index", "all". kind: Filter by classification — "primitive", "struct", "enum", "data_interface", "object", or "all" (default). filter: Case-insensitive substring match on type name. For data interfaces also matches against the short name (without the "NiagaraDataInterface" prefix), so filter="ArrayFloat" finds NiagaraDataInterfaceArrayFloat. max_results: Cap on returned entries (default 500). Returns entries with name, kind, size_bytes, plus: - data_interface: class_path, short_name, category, description - enum: enum_path, num_entries - struct: struct_path, description

List all scratch pad modules on a Niagara System. Returns name, path, usage, node counts, and custom HLSL node count per script.

List input + output parameters of a Niagara script. Outputs come from UNiagaraNodeOutput::Outputs (the Output Dynamic Input node for dynamic inputs, or Output Module for modules). Inputs come from graph script variable metadata filtered to Module.* / Input.* / User.* namespaces. Useful for discovering what an "Add Parameter" dropdown would show or for diff-auditing a graph before/after mutation.

List Niagara System assets in the project. Args: path: Content Browser directory to search in name_filter: Optional substring filter on asset name max_results: Maximum number of results to return

List valid enum values for StateTree properties. Use this to discover valid values BEFORE creating states, transitions, or conditions. Does not require an asset_path — these are engine-level constants. Categories: - "trigger": OnStateCompleted, OnStateSucceeded, OnStateFailed, OnTick, OnEvent - "priority": Low, Normal, Medium, High, Critical - "state_type": State, Group, Linked, LinkedAsset, Subtree - "selection_behavior": None, TryEnterState, TrySelectChildrenInOrder, etc. - "operand": And, Or (for condition expressions) Args: category: One of: trigger, priority, state_type, selection_behavior, operand

List available task/evaluator/condition types for the StateTree's schema. Args: asset_path: Content path to the StateTree asset (determines schema) category: "task", "evaluator", or "condition" filter: Optional substring filter on class name

List all available UInputTrigger subclasses with their editable properties. Useful for discovering valid trigger_type values. Returns name, display_name, parent class, and editable property names/types.

List available UWidget subclasses. Use filter to narrow results.

Move a material expression node to a new graph position.

Move a widget to a different parent panel. Cannot move root or into own descendants.

Open an asset in the UE5 editor (blueprint, material, data table, widget, etc.).

Query profiling data from a loaded .utrace trace. RECOMMENDED WORKFLOW (compact output, saves context): 1. "diagnose" — ONE CALL full report: verdict, findings, categories, top timers 2. "spikes" — Auto worst frames + category breakdown 3. "flame" — Top timers by exclusive (self) time — actual bottleneck code 4. "hotpath" — Drill into a category or event's children 5. "search" — Find a timer across all frames (stats + worst frames) 6. "histogram" — Frame time distribution (periodic hitches vs sustained?) 7. "compare" — Compare frame vs trace median, show outlier events Use performance_start_trace/stop_trace to record and auto-load a trace, or use query="load" to load an existing .utrace file. Args: query: Type of analysis. One of: SMART QUERIES (compact output, do analysis in C++): - "diagnose" — ONE CALL full performance report. Returns: verdict string, severity-rated findings with recommendations, category breakdown (avg per-frame exclusive ms), top 10 exclusive timers, GPU-bound detection. Start here. - "bottlenecks" — Auto-categorize frame into Animation/Slate/Network/etc with total time per category and top event. VERY compact. - "hotpath" — Drill into a category or event's children, sorted by time - "compare" — Compare frame vs trace median, show outlier events only - "spikes" — Auto worst frames + category breakdown (combines worst_frames + bottlenecks). Shows top 3 categories per spike frame. - "search" — Find a timer across ALL frames. Returns min/avg/max/p95/p99 stats + worst frames list. Use filter= for timer name. - "histogram" — Frame time distribution. Shows bucket counts + budget summary (on-budget / slightly over / 2x / 4x over). - "flame" — Top timers by EXCLUSIVE (self) time. Shows per-frame avg exclusive ms, self_pct (excl/incl ratio), category label. Answers "what code actually consumes the most CPU?" STANDARD QUERIES: - "load" — Load an existing .utrace file (requires trace_path) - "summary" — Overview: duration, frame count, avg/min/max/p95 frame time - "worst_frames" — Find the slowest frames with top events - "frame_details" — Per-thread timing breakdown for one frame (use bottlenecks first!) - "timer_stats" — Aggregated stats for timers (top N by inclusive time) - "butterfly" — Callers/callees for a specific timer (requires timer_name) - "threads" — List all threads in the trace - "counters" — List counters (GPU memory, draw calls, etc.) PROVIDER QUERIES: - "net_stats" — Network profiling (needs 'net' channel) - "loading" — Asset loading analysis (needs 'loadtime' channel) - "logs" — Log messages (needs 'log' channel) - "memory" — LLM memory tags (needs 'memtag' channel) - "regions" — Timing regions (needs 'region' channel) - "bookmarks" — Bookmark events (needs 'bookmark' channel) EXTENDED PROVIDER QUERIES: - "session" — Session metadata: platform, app, project, build version, branch, changelist, config, target type. Essential context. - "modules" — Loaded modules/DLLs with symbol resolution stats. Shows resolved/failed/pending counts per module. - "file_io" — File I/O activity: per-file read/write counts, bytes, duration. Sorted by impact. Needs 'loadtime' channel. - "tasks" — Task graph: task lifecycle (created→finished), duration, wait time, prerequisites, nested tasks. Sorted by duration. - "context_switches" — CPU core scheduling: core count, per-thread context switch frequency. Needs elevated/admin trace. - "allocations" — Memory allocation timeline: peak/min memory, live alloc count, alloc/free events, heaps, swap. Needs 'memalloc'. - "stack_samples" — CPU sampling profiler: flat profile of hotspot functions by sample count with percentage. Needs sampling enabled. - "screenshots" — Screenshot metadata captured during trace recording. Needs 'screenshot' channel. trace_path: (load) Absolute path to a .utrace file frame_index: (bottlenecks, hotpath, compare, frame_details) Which frame (0-based) target_fps: (bottlenecks, spikes, histogram) Target FPS for budget (default: 60) category: (hotpath) Category to drill into: Animation, Slate, Network, Physics, Rendering, Gameplay, Audio, Loading, GarbageCollection, Other event_name: (hotpath) Specific event name to show children of min_deviation_pct: (compare) Min deviation % to report (default: 50) thread: (bottlenecks, hotpath, compare, spikes, search) Thread (default: GameThread) count: (worst_frames, spikes, search, timer_stats, hotpath, file_io, tasks, modules, stack_samples) Max results threshold_ms: (worst_frames, spikes) Only show frames slower than this max_depth: (frame_details, butterfly, hotpath) Max call stack depth min_duration_ms: (frame_details) Hide events shorter than this thread_name: (frame_details) Filter to threads containing this string filter: (search, timer_stats, counters, logs, memory, regions, bookmarks, modules, file_io, tasks, stack_samples) Name/path filter start_time: Start of time range in seconds (-1 = trace start) end_time: End of time range in seconds (-1 = trace end) include_values: (counters) Include sampled values max_samples: (counters) Max value samples per counter timer_name: (butterfly) Function/scope name to analyze mode: (butterfly) "callers", "callees", or "both" connection_index: (net_stats) Connection index for packet details verbosity: (logs) Filter: Fatal, Error, Warning, Display, Log, Verbose tracker: (memory) Memory tracker ID bucket_size_ms: (histogram) Custom bucket width in ms (0 = auto based on target_fps)

Start recording a performance trace from the running Unreal Editor. Records CPU, GPU, frame timing, and other profiling data to a .utrace file. Call performance_stop_trace when done to stop recording and auto-load for analysis. Args: file_path: Where to save the .utrace file. Empty = auto-generate in Saved/Profiling/ channels: Comma-separated trace channels or preset name. Presets: "default" (cpu,gpu,frame,log,bookmark,screenshot,region) Individual: cpu, gpu, frame, log, bookmark, screenshot, region, memtag, memalloc For network profiling add: net For asset loading add: loadtime For memory tracking add: memtag Example: "default,net,loadtime,memtag" for full profiling

Stop the active trace recording and optionally load it for analysis. After stopping, the trace is automatically loaded so you can immediately query it with performance_analyze_insight. Args: auto_load: If True (default), auto-load the recorded trace for analysis

Read complete Blueprint content: event graph, functions, variables, components.

Force recompile and save a material. Reports success/failure.

Remove a modifier from an Input Action by index. Use get_input_action to see current modifier indices.

Remove a trigger from an Input Action by index. Use get_input_action to see current trigger indices.

Remove a key mapping from an Input Mapping Context. Remove by index OR by action_path + key combination. Args: context_path: Path to the IMC asset index: Mapping index to remove (from get_input_mapping_context) action_path: Alternative — remove by action + key combo key: Alternative — remove by action + key combo