> For the complete documentation index, see [llms.txt](https://animotive.gitbook.io/animotive-kimodo/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://animotive.gitbook.io/animotive-kimodo/working-in-sequencer/animotive-kimodo-track.md).

# Animotive Kimodo Track

The Animotive Kimodo track is the day-to-day authoring surface of the plugin. It lives on a body skeletal-mesh binding in Sequencer, holds one or more prompt sections, and — together with its companion **Kimodo Constraints** track — describes the generation that runs when you click **Generate with Kimodo**.

For the companion track, see [Constraints Track](/animotive-kimodo/working-in-sequencer/constraints-track.md). For the first-time walkthrough, see [Prompting with Animotive Kimodo](/animotive-kimodo/getting-started/prompting-with-animotive-kimodo.md).

## What it is

The Animotive Kimodo track is a custom Sequencer track that attaches to a body skeletal-mesh binding. It carries:

* One or more **prompt sections**, each spanning a time range and carrying a single required text prompt.
* A **Forward Retargeter** slot (Soma → Character) and a **Reverse Retargeter** slot (Character → Soma).

Adding the track also creates a sibling **Kimodo Constraints** track on the same binding in the same undo step. The two tracks are designed to be used together: the Animotive Kimodo track describes *what* happens between key poses; the Constraints track marks *which* body parts are pinned at which frames.

## Adding the track

1. In the Sequencer outliner, right-click the body skeletal-mesh sub-binding (typically named `Body` on a MetaHuman).
2. Choose **Add Animotive Kimodo Track**.

The menu entry only appears on bindings whose skeleton has a `pelvis` or `root` bone — face rigs and prop attachments are filtered out.

> **One track per binding.** The **Add Animotive Kimodo Track** entry disappears from a binding's menu only once the *full* pair is present — both the Animotive Kimodo track and its **Kimodo Constraints** sibling. There is no standalone "Add Constraints Track" entry. Deleting the Animotive Kimodo track does not auto-delete the Constraints sibling, or vice versa — remove the other separately if you want both gone.
>
> **Re-adding a deleted companion.** If you delete just one of the pair, the **Add Animotive Kimodo Track** entry reappears on that binding. Choosing it again is self-healing: it restores whichever track is missing in a single undo step and leaves the surviving track — including its prompt sections and retargeter slots — untouched.
>
> **One character per click.** Each **Generate with Kimodo** click generates for the character whose Animotive Kimodo binding is selected. Multi-character scenes work in a single sequence — select each character in turn and click **Generate with Kimodo** for each. They generate one after another rather than simultaneously.

For non-MetaHuman characters that need a wizard pass first, see [First-Time Character Setup](/animotive-kimodo/setup/first-character.md).

## Prompt sections

A prompt section is a contiguous range on the track's row with a single text prompt describing the motion to generate within that range.

A prompt is **required** on every section, but an empty section does not block you up front. When you click **Generate with Kimodo**, the plugin pops the prompt-entry dialog for each empty section in turn — titled **Prompt for section N of M (frames Lo–Hi)** so you know which one you are filling — and writes what you type back into the section (one undo step). Cancelling the dialog, or submitting it empty, cancels the whole export.

To author a prompt:

1. Right-click the section (or double-click it) and choose **Edit Prompt…**
2. The dialog opens with the title **Animotive Kimodo Prompt** and the body label **Describe the motion you want Kimodo to generate.**
3. Type a short description focused on the action — `walks confidently`, `crouches and reaches forward`, `pivots and waves`. Keep prompts focused on motion; the model does not need scene-setting context.

> **Descriptors and emotions work.** Adverbs and emotional qualifiers shape the result: `tiredly walks forward`, `cautiously crouches`, `angrily waves`, `nervously paces`. Combining a verb with a descriptor is usually more useful than a longer scene description — start with one action and one qualifier, then iterate.

4. Click **Submit** to save, or **Cancel** to discard.

A track can carry multiple prompt sections side by side; each describes its own time range. At export, every section across every Animotive Kimodo track in the focused sequence is collapsed into one Kimodo generation, stitched together in timeline order. Adjacent sections cannot overlap — **Add Section** in the right-click menu clamps the new section to fit between neighbours.

## Retargeter slots

The track's details panel exposes two retargeter slots:

* **Forward Retargeter** — Soma → Character. Applied when Kimodo's returned BVH is retargeted onto your character.
* **Reverse Retargeter** — Character → Soma. Used by the retarget-first export path when sampling your character's posed motion into Kimodo's input space.

Both slots are auto-detected when the track is first added. For a MetaHuman this resolves to the retargeter pair that ships with the plugin; for a character set up via the Retarget Wizard, it resolves to the pair under `/Game/Kimodo/Generated/<MeshName>/`.

When a slot is not resolved, a warning chip appears on the track row. The tooltip names which slot is missing. To recover, right-click the track and choose **Re-Detect Retargeters**. If detection still finds no match, run the Retarget Wizard for the character (see [First-Time Character Setup](/animotive-kimodo/setup/first-character.md)), then re-detect. As a last resort, assign the retargeter assets manually in the details panel.

## Right-click context menu

| Entry                     | What it does                                                                                                               |
| ------------------------- | -------------------------------------------------------------------------------------------------------------------------- |
| **Add Section**           | Adds a new prompt section (about 2 seconds long) at the playhead, clamped between any neighbouring sections.               |
| **Edit Prompt…**          | Opens the **Animotive Kimodo Prompt** dialog for the section under the cursor. Same action as double-clicking the section. |
| **Re-Detect Retargeters** | Re-runs the project-wide scan that fills the Forward and Reverse slots.                                                    |

## Advanced

These settings are tuneable but do not need attention for typical use. The defaults work for most generations.

### Transition frames

The track-level property **Num Transition Frames** (default `5`) controls how Kimodo blends adjacent prompt sections. It only applies when the track has two or more sections — with a single section there is nothing to blend.

Raise it (toward 60) when adjacent prompts describe physically distant motions (e.g. `running` → `sitting down`) and a hard cut would look jarring. Lower it (toward 0) to keep each section more responsive to its own prompt, at the cost of a sharper boundary. The value is in the track's details panel.

### Seed (Fixed vs Random)

The track-level property **Seed Mode** controls the random-number-generator state Kimodo uses when solving in-betweens. Two settings of the same keyed sequence with the same seed produce identical motion; different seeds explore the solver's distribution.

* **Fixed** (default): use the **Fixed Seed** integer (default `24`) verbatim. Reproducible — re-running on the same scene yields the same motion.
* **Random**: the plugin samples a fresh non-negative `int32` per export. The actual integer used is captured into the [Kimodo Monitor](/animotive-kimodo/working-in-sequencer/monitor-panel.md) row (visible in the tooltip) and the editor log, so you can copy it back into **Fixed Seed** to reproduce a Random run later.

The **Fixed Seed** numeric field is greyed out automatically when **Seed Mode = Random** — the resolved random value is not stored back into the property; check the Monitor or `Saved/Logs/Kimodo.log` for the seed used in any given generation.

### Auto-fill from Control Rig keyframes

If the binding already has Control Rig keyframes laid down on the timeline, the plugin can scaffold prompt sections from those keyframe times in one action:

1. Right-click the Animotive Kimodo track row.
2. Choose **Add Prompts Between Control Rig Keyframes**.

One empty-prompt section is created per gap between adjacent keyframes (plus leading and trailing gaps to the playback range). Each section starts with no prompt — author the text via **Edit Prompt…**, or leave them and fill each one when the export-time dialog prompts you (see [Prompt sections](#prompt-sections)). If the track already has sections, a confirmation dialog appears before they are replaced.

### Control-Rig-driven constraint export

The track property `Use Control Rig Constraints` (default on) controls whether the binding's Control Rig keyframes are sampled and emitted as constraints alongside the prompts. Turning it off makes the run a pure text-to-anim generation over the section ranges, independent of any Control Rig keys on the binding.

## See also

* [Constraints Track](/animotive-kimodo/working-in-sequencer/constraints-track.md)
* [Keying Poses](/animotive-kimodo/working-in-sequencer/keying-poses.md)
* [Animotive Kimodo Monitor Panel](/animotive-kimodo/working-in-sequencer/monitor-panel.md)
* [Import & Results](/animotive-kimodo/working-in-sequencer/import-and-results.md)
* [First-Time Character Setup](/animotive-kimodo/setup/first-character.md)