Character rig support

Which character rigs work with the Convai Unreal Engine plugin, what setup each requires, and where to find the rig-specific setup guides.

The Convai Unreal Engine plugin drives facial animation through the UConvaiFaceSyncComponent, which outputs facial animation data controlled by the EC_LipSyncMode enum. Blendshape modes output blendshape channels; VisemeBased mode outputs viseme values. The lip sync mode you select determines which rig types are compatible.

Supported rigs

Rig type	Lip sync mode	Out-of-the-box	Additional setup
MetaHuman	MetaHuman Blendshapes (BS_MHA)	Yes; this is the default mode	Bundled Convai_MetaHuman_BodyAnim and Convai_MetaHuman_FaceAnim animation Blueprints included
Reallusion Character Creator (CC4 and CC5)	CC5: MetaHuman Blendshapes (BS_MHA); CC4: ARKit Blendshapes (BS_ARKit) or CC4 Extended Blendshapes (BS_CC4_Extended)	No	Follow the Reallusion setup guide for animation Blueprint setup
ARKit-compatible rigs	ARKit Blendshapes (BS_ARKit)	No	Configure LipSyncMode to BS_ARKit; wire 61 blendshape channels to your rig
Viseme-based rigs	Viseme Based (VisemeBased)	No	Configure LipSyncMode to VisemeBased; map 15 OVR visemes to your rig
Single-mesh generic rigs	Auto or Off	Partial	No facial animation without a compatible mode configured

MetaHuman characters

MetaHuman is the default rig target. The plugin ships with Convai_MetaHuman_BodyAnim and Convai_MetaHuman_FaceAnim under Content/MetaHumans/Animations/, and sets LipSyncMode to BS_MHA by default on UConvaiFaceSyncComponent. No separate animation Blueprint download is required for MetaHuman setup.

The GetBodyAndFaceSkeletalMeshComponents utility function (Blueprint-callable, category Convai|Utilities) resolves the body and face Skeletal Mesh Component references on a MetaHuman-style Actor automatically using bone-name heuristics. MetaHuman rigs that separate the face mesh from the body mesh are handled correctly.

Use the MetaHuman setup guide for the rig-specific setup flow.

Set up a MetaHuman character

MetaHuman character setup walkthrough

Reallusion Character Creator (CC4 and CC5)

Reallusion Character Creator CC4 and CC5 characters use a separate animation Blueprint workflow from the MetaHuman setup. The plugin includes FixCC5LipsyncPostProcessBlendshapes (Blueprint-callable, category Convai|LipSync) to apply a post-process fix for CC5 rigs that use ExpBoneData in their post-process animation instance.

Reallusion setup uses a separate Convai animation Blueprint. Follow the Reallusion setup guide before assigning a CC4 or CC5 character to your scene.

After the Reallusion animation Blueprint is assigned, set LipSyncMode on the UConvaiFaceSyncComponent attached to your CC character based on the rig generation: use MetaHuman Blendshapes (BS_MHA) for CC5, and ARKit Blendshapes (BS_ARKit) for CC4. For CC4 rigs that drive the extended Reallusion blendshape set, use CC4 Extended Blendshapes (BS_CC4_Extended) instead.

Use the Reallusion setup guide for the download and animation Blueprint assignment steps.

Set up a Reallusion (CC) character

Reallusion CC character setup walkthrough

ARKit-compatible rigs

Any rig that exposes ARKit-style blendshape targets can use BS_ARKit mode. The plugin drives 61 blendshape channels in this mode: the 52 standard ARKit face targets plus 9 head and eye rotation channels (HeadYaw, HeadPitch, HeadRoll, LeftEyeYaw, LeftEyePitch, LeftEyeRoll, RightEyeYaw, RightEyePitch, RightEyeRoll). Set LipSyncMode to ARKit Blendshapes on the UConvaiFaceSyncComponent and wire the blendshape output to your rig's animation Blueprint.

Viseme-based rigs

VisemeBased mode outputs 15 OVR viseme values (named sil, PP, FF, TH, DD, kk, CH, SS, nn, RR, aa, E, ih, oh, ou). Any rig that maps those 15 viseme names to morph targets or bone poses can use this mode.

Rig detection utility

GetBodyAndFaceSkeletalMeshComponents inspects an Actor's Skeletal Mesh Component references and resolves the body and face meshes using bone-name heuristics. It handles:

• Actors with a single skeletal mesh — the mesh becomes the body, face output is null.
• Actors with two or more skeletal meshes — each mesh is scored against body token bones (pelvis, hand, neck, and others) and face token bones (eye, jaw, brow, and others); the highest-scoring mesh wins each role.
• Actors with no skeletal mesh components — both outputs are null.

This function is intended for MetaHuman-style rigs but also handles single-mesh and unrecognized rigs without crashing.

Next steps

After confirming rig support, check the network requirements before installing or validating the plugin.

Network and API requirements