CodeFizz
ToolsDemoDocsRoadmapChangelogPricingFAQ
Get the plugin

Getting Started

  • Introduction
  • Quick Start
  • Requirements
  • Installing the CLI
  • License & Machines

Guides

  • Installing the Plugin
  • The AI Skill
  • CLI vs MCP
  • Multiple Editors
  • How It Works
  • Discovering Commands

Walkthroughs

  • Build a Blueprint with AI
  • Author a Material Graph
  • Create a Niagara System
  • Rig a Character with AI
  • Build an Animation Blueprint

Reference

  • Reference
  • Control Rig
  • Niagara
  • PCG
  • Materials
  • StateTree
  • Sequencer
  • Level Actors
  • Project Settings
  • Blueprints
  • Blueprint Structs
  • Enhanced Input
  • Asset Management
  • Bulk Asset Ops
  • Data Assets
  • Object Properties
  • UMG Widgets
  • Data Tables
  • Curves
  • Animation
  • Console Commands
  • Profiling
  • Core
  • Debug

Help

  • FAQ
  • Troubleshooting
  • For AI Agents
DocsWalkthroughsBuild an Animation Blueprint

Build an Animation Blueprint

A guided example: build a locomotion blend space, an Animation Blueprint with a state machine, bound transition rules, an upper body layer, and a montage with notifies, then validate it.

Loading…
PreviousRig a Character with AINextReference
CodeFizz

A drop-in Unreal Engine 5 plugin that exposes the entire editor surface (Blueprints, Materials, Niagara, PCG, StateTree, Control Rig, Insights profiling, and more) over the Model Context Protocol. Connect Claude Code, Cursor, VS Code, or any MCP-compatible client and let your AI build inside the engine.

Product

  • Features
  • Docs
  • Tools
  • Demo
  • Roadmap
  • Changelog
  • Pricing
  • FAQ

Resources

  • Install guide
  • Discord
  • YouTube
  • Open-source edition
  • Manage subscription

Legal

  • Refund Policy
  • Privacy Policy
  • Terms of Service

© 2026 CodeFizz. All rights reserved.

CodeFizz Editor Agent is a CodeFizz product. Payments processed by Polar Software, Inc. (Polar), the Merchant of Record.

This walkthrough builds a working locomotion setup from nothing: a blend space for ground movement, an Animation Blueprint, a state machine with real transition rules, an upper body layer, and a montage with notifies. Every command in the family is listed in the Animation reference.

1

Build the locomotion blend space

A 2D blend space with direction across and speed up. Name and range each axis, then drop a sample per animation.

bash
cfa create_blend_space --name BS_Locomotion --path /Game/Anims --skeleton /Game/Chars/SK_Hero --kind blendspace_2d
2

Range the axes

Direction runs -180 to 180 and wraps, speed runs 0 to your jog speed. Do this before adding samples so the samples land where you expect.

bash
cfa set_blend_space_axis --blend-space-path /Game/Anims/BS_Locomotion --axis x --display-name Direction --min -180 --max 180
3

Place the samples

One call per animation. An idle row at speed 0, a walk ring, a jog ring. The blend space triangulates them for you.

bash
cfa add_blend_space_sample --blend-space-path /Game/Anims/BS_Locomotion --animation /Game/Anims/Jog_Fwd --x 0 --y 500
4

Create the Animation Blueprint

Bound to a skeleton, with a preview mesh so you can see it move. It arrives with an AnimGraph ready to use.

bash
cfa create_anim_blueprint --name ABP_Hero --path /Game/Anims --skeleton /Game/Chars/SK_Hero --preview-mesh /Game/Chars/SKM_Hero
5

Add the state machine and its states

States, plus conduits for shared entry logic and aliases that stand in for several states at once.

bash
cfa add_anim_state --blueprint-path /Game/Anims/ABP_Hero --graph AnimGraph/Locomotion --name Idle
6

Wire the transitions and bind their rules

Bind a bool variable as the rule. Add --rule-inverted for the way back, so leaving a state does not need a hand built NOT.

bash
cfa add_anim_transition --blueprint-path /Game/Anims/ABP_Hero --graph AnimGraph/Locomotion --from Idle --to Move --rule-variable bIsAccelerating --crossfade-duration 0.15
7

Add an upper body layer

Put aiming or a weapon pose on its own layer, then blend it over locomotion per bone. Share it between blueprints with a layer interface.

bash
cfa add_anim_layer --blueprint-path /Game/Anims/ABP_Hero --name UpperBodyPose
8

Build a montage with notifies

Assemble the montage, add named sections you can chain, then place notifies and notify states on their own tracks.

bash
cfa create_anim_montage --name AM_Attack --path /Game/Anims --skeleton /Game/Chars/SK_Hero --animation /Game/Anims/Attack_01
9

Tidy the graph

Lay the state machine out as a readable block, with each transition sitting between the two states it joins.

bash
cfa arrange_blueprint_graph --blueprint-path /Game/Anims/ABP_Hero --function-name AnimGraph/Locomotion
10

Validate before you play

Returns a verdict plus every compiler warning, including transitions that can never fire because their rule was never bound.

bash
cfa validate_anim_blueprint --blueprint-path /Game/Anims/ABP_Hero
A transition with no rule can never be taken
Adding a transition without --rule-variable leaves its Can Enter Transition pin empty, so the state machine reaches that state and never leaves it. The blueprint still compiles, so nothing stops you. Run validate_anim_blueprint and read the warnings: it names every transition that can never fire. For a time based transition set bAutomaticRuleBasedOnSequencePlayerInState instead, and clear Loop Animation on that state's sequence, because a looping animation never ends and the rule never fires.
Nested graphs are addressed by a slash path
Everything inside an Animation Blueprint is a graph, so the ordinary Blueprint graph commands work on them once you address them correctly: AnimGraph/Locomotion/Idle. Use list_anim_graphs to see the tree. Transition rule graphs are all named Transition, so a bare name is ambiguous and refused; pass the #N suffix the listing gives you.
Compile a layer interface before you implement it
A layer only reaches the implementing blueprint once it exists as a function on the interface class, which compiling is what creates. Skip that and the interface is listed but the layer graph is silently absent. It recovers: compile the interface, then the implementer. Exact parameters for any command are in the Animation reference, or run cfa describe add_anim_transition.