> 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/going-further/recipes.md).

# Recipes

Task-focused walkthroughs for common authoring jobs with the Animotive Kimodo plugin. Each recipe assumes the plugin is installed, a backend is configured, and you have completed at least one end-to-end generation. If not, start with [Quick Start](/animotive-kimodo/getting-started/quick-start.md) and [Prompting with Animotive Kimodo](/animotive-kimodo/getting-started/prompting-with-animotive-kimodo.md).

The recipes are independent — read the ones that match the job in front of you.

| Recipe                                                                          | Use when                                                                  |
| ------------------------------------------------------------------------------- | ------------------------------------------------------------------------- |
| [Single-shot in-between](#single-shot-in-between-two-key-poses)                 | You need a short clip between two authored poses, fast.                   |
| [Iterative prompt tweaking](#iterative-prompt-tweaking)                         | You want to explore prompt variants against the same key poses.           |
| [Foot-planted walk](#foot-planted-walk)                                         | You need a locomotion clip with feet staying on the ground.               |
| [Text-to-anim (no key poses)](#text-to-anim-no-key-poses)                       | You want a result from a prompt alone, with no spatial control.           |
| [Rough scene blocking with pelvis only](#rough-scene-blocking-with-pelvis-only) | You are staging a character's path before committing to a full pose pass. |
| [Multi-character scene](#multi-character-scene)                                 | The shot involves two or more characters generating motion.               |
| [Bake to Control Rig for hand polish](#bake-to-control-rig-for-hand-polish)     | You want frame-by-frame control over an injected animation.               |
| [Reproduce a generation by seed](#reproduce-a-generation-by-seed)               | You liked a Random-mode result and want to lock it in or re-run it.       |

***

## Single-shot in-between (two key poses)

**Goal:** the fastest path from two authored poses to a 1-3 second in-between clip.

1. Open the Level Sequence containing your character in Sequencer.
2. Right-click the body sub-binding (typically `Body`) and choose **Add Animotive Kimodo Track** if the track is not already present.
3. Scrub to the start frame, pose the character on its Control Rig, press **S** to key. Scrub forward 30-90 frames, pose differently, press **S** again.
4. Right-click the empty section on the Animotive Kimodo track and choose **Edit Prompt…**. Type one short action description (`walks confidently`, `crouches and reaches`). Click **Submit**.
5. Click **Generate with Kimodo** in the Sequencer toolbar.
6. Watch the row in the **Kimodo Monitor** advance to **Completed**.
7. Scrub the injected section on the timeline.

**Look for:** the injected skeletal animation section appears on the binding and the original Control Rig section is automatically muted over the same range.

**Variant:** this is the same loop covered in [Prompting with Animotive Kimodo](/animotive-kimodo/getting-started/prompting-with-animotive-kimodo.md); this recipe is the shorthand.

*See also:* [*Iterative prompt tweaking*](#iterative-prompt-tweaking)*.*

***

## Iterative prompt tweaking

**Goal:** explore prompt variants against the same start and end poses.

1. Start from a sequence that already produced one good generation (key poses keyed, prompt section authored).
2. Leave the two Control Rig keyframes untouched.
3. Right-click the prompt section on the Animotive Kimodo track and choose **Edit Prompt…**. Edit the text. Click **Submit**.
4. Click **Generate with Kimodo**.
5. Wait for the new row to reach **Completed** in the **Kimodo Monitor**.
6. Compare results: the new injected section stacks on top of the previous one and auto-mutes it over the overlapping range, so the latest plays by default. To A/B, un-mute the older section (and mute the newer) one at a time.
7. Once you have a favourite, delete the losing injected sections and the underlying Anim Sequence assets under `/Game/Kimodo/Generated/<MeshName>/Sequences/`.

**Look for:** each click produces a distinct time-stamped Anim Sequence asset — useful for A/B comparison, and harmless to keep around until you decide.

> **Descriptors and emotions shape the result.** Adverbs and emotional qualifiers carry weight in the prompt. Compare `walks forward` to `tiredly walks forward`, `confidently walks forward`, `cautiously walks forward`. Iterating on a single descriptor at a time is a fast way to dial in performance without re-keying poses.

*See also:* [*Single-shot in-between*](#single-shot-in-between-two-key-poses)*,* [*Import & Results*](/animotive-kimodo/working-in-sequencer/import-and-results.md)*.*

***

## Foot-planted walk

**Goal:** a multi-stride locomotion clip with feet planted on the ground.

1. On the body binding, key the character's Control Rig at each stride apex along the timeline — start, mid-stride, mid-stride, end. At each keyframe, position the planting foot exactly where it should make contact with the ground.
2. If not already present, right-click the body binding and choose **Add Animotive Kimodo Track** (the **Kimodo Constraints** sibling appears alongside it).
3. Expand the **Kimodo Constraints** track in the outliner to reveal its six channel rows.
4. Scrub to the frame where the left foot first plants, select the **Left Foot** channel row, press **Enter** to add a trigger key.
5. Scrub to the next plant frame, select the **Right Foot** channel row, press **Enter**. Continue alternating across the stride pattern.
6. Right-click the prompt section on the Animotive Kimodo track and author the walk prompt (`walks forward at a steady pace`, `strides confidently`). Click **Submit**.
7. Click **Generate with Kimodo**.
8. On the injected result, scrub through the constrained frames and confirm each planted foot stays in place across that frame.

**Look for:** at each constrained frame, the foot sits exactly where the Control Rig posed it — the generator solves the rest of the body to match. If a foot appears to slip, the most common cause is the foot not being posed at the planting location at that frame.

*See also:* [*Constraints Track*](/animotive-kimodo/working-in-sequencer/constraints-track.md)*,* [*Keying Poses*](/animotive-kimodo/working-in-sequencer/keying-poses.md)*.*

***

## Text-to-anim (no key poses)

**Goal:** generate a motion from a prompt alone, with no spatial control over start or end pose.

1. Right-click the body sub-binding and choose **Add Animotive Kimodo Track** if not already present.
2. Do not author any Control Rig keys. If the binding already carries Control Rig keys you want to ignore, select the Animotive Kimodo track and toggle the **Use Control Rig Constraints** property off in the details panel so stray keys do not feed into the export.
3. Right-click the empty section on the Animotive Kimodo track, choose **Edit Prompt…**, and author the motion description.
4. Drag the section's edges to cover the time window you want generated.
5. Click **Generate with Kimodo**.
6. Wait for **Completed**, then scrub the injected section.

**Look for:** the lowest-friction way to get a result. The trade-off is no control over start or end pose — the model picks both. If you have a hard requirement on either end, add a Control Rig keyframe at that frame and re-enable **Use Control Rig Constraints**.

*See also:* [*Single-shot in-between*](#single-shot-in-between-two-key-poses)*,* [*Animotive Kimodo Track*](/animotive-kimodo/working-in-sequencer/animotive-kimodo-track.md)*.*

***

## Rough scene blocking with pelvis only

**Goal:** lay out a character's path through the level by keying only the pelvis on the Control Rig — no full pose required.

1. Scrub the playhead to the start frame. Click the character in the viewport to reveal Control Rig handles. Grab only the pelvis handle and place it at the start location. Press **S** to key.
2. Scrub to the midpoint of the desired motion. Move the pelvis handle to the next staging location. Press **S**.
3. Scrub to the end frame. Move the pelvis to the final location. Press **S**.
4. Right-click the body binding and choose **Add Animotive Kimodo Track** if not already present.
5. Right-click the prompt section, choose **Edit Prompt…**, and author a descriptive prompt covering the action (`walks across the room and stops`, `paces between two marks`). Click **Submit**.
6. Click **Generate with Kimodo**.
7. Review the result on the timeline.

**Look for:** a Control Rig keyframe that affects the pelvis automatically registers as a **2D Root** sample at that frame — no manual Constraints keying required. The upper body stays free for the prompt to drive, while the floor footprint matches your staging.

*See also:* [*Constraints Track*](/animotive-kimodo/working-in-sequencer/constraints-track.md)*,* [*Keying Poses*](/animotive-kimodo/working-in-sequencer/keying-poses.md)*.*

***

## Multi-character scene

**Goal:** author a shot with two or more characters in the same sequence.

Multiple characters can live in one Level Sequence. Each **Generate with Kimodo** click generates for the currently selected character, so you generate them one at a time rather than simultaneously.

1. Add each character to the same Level Sequence, and add an **Animotive Kimodo Track** to each one's body binding.
2. Author each character independently — key poses, prompt sections, and any constraint keys on its own tracks.
3. Select the first character's Animotive Kimodo binding in the Sequencer outliner.
4. Click **Generate with Kimodo**. Wait for **Completed** in the **Kimodo Monitor**.
5. Select the next character's binding and click **Generate with Kimodo** again. Repeat for each character.
6. Scrub the sequence — each character plays its generated animation in the shared timeline.

**Look for:** the export resolves the character whose Animotive Kimodo binding is selected. If a generation lands on the wrong character, confirm the correct binding (and its sibling Control Rig) is selected before clicking **Generate with Kimodo**.

*See also:* [*Animotive Kimodo Track*](/animotive-kimodo/working-in-sequencer/animotive-kimodo-track.md)*.*

***

## Bake to Control Rig for hand polish

**Goal:** convert an injected Anim Sequence back into Control Rig keyframes for frame-by-frame tweaking.

1. Wait for a generation to reach **Completed** in the **Kimodo Monitor** and confirm the injected skeletal animation section is on your body binding in Sequencer.
2. Right-click the injected skeletal animation section in the timeline.
3. Choose **Bake to Control Rig** from the context menu.
4. Unreal generates a new Control Rig section on the binding with keys baked from the Anim Sequence. Scrub through it to confirm the motion matches the source clip.
5. Refine the baked Control Rig keys — adjust hands, head turns, contact frames, anything you want manual control over.
6. Save the sequence.

**Look for:** this is a one-way conversion within the section. Once baked, the section is a Control Rig section and is no longer a reference to the Anim Sequence asset. The original Anim Sequence remains in the Content Browser under `/Game/Kimodo/Generated/<MeshName>/Sequences/` as a backup — keep it until you are happy with the polished result.

*See also:* [*Import & Results*](/animotive-kimodo/working-in-sequencer/import-and-results.md)*,* [*Keying Poses*](/animotive-kimodo/working-in-sequencer/keying-poses.md)*.*

***

## Reproduce a generation by seed

**Goal:** re-run a result you liked, or lock a good Random-mode take so it stops changing.

The **seed** drives the solver's random state. Same key poses + same prompt + same seed → identical motion. Two settings on the Animotive Kimodo track control it: **Seed Mode** (Fixed / Random) and **Fixed Seed**.

1. Generate with **Seed Mode = Random** while you explore — each run gets a fresh seed and a different take.
2. When a result is good, find its row in the [Kimodo Monitor](/animotive-kimodo/working-in-sequencer/monitor-panel.md). The row shows `seed N` — the exact integer that produced it.
3. On the Animotive Kimodo track, set **Seed Mode = Fixed** and type that `N` into **Fixed Seed**.
4. Re-export. The generation now reproduces the same motion every time.

**Look for:** in Random mode the **Fixed Seed** field is greyed out — the sampled value is not written back to the property, so the Monitor row (or `Saved/Logs/Kimodo.log`, line `OnExportToKimodo: seed=N`) is the only record of which seed ran. Capture it before clearing history.

*See also:* [*Animotive Kimodo Track*](/animotive-kimodo/working-in-sequencer/animotive-kimodo-track.md)*,* [*Monitor Panel*](/animotive-kimodo/working-in-sequencer/monitor-panel.md)*.*

***

## See also

* [Animotive Kimodo Track](/animotive-kimodo/working-in-sequencer/animotive-kimodo-track.md) — the parent authoring surface for prompts and time ranges.
* [Constraints Track](/animotive-kimodo/working-in-sequencer/constraints-track.md) — the companion track for pinning body regions at specific frames.
* [Troubleshooting](/animotive-kimodo/going-further/troubleshooting.md) — symptom-driven fixes for the issues most likely to interrupt a recipe.
* [Prompting with Animotive Kimodo](/animotive-kimodo/getting-started/prompting-with-animotive-kimodo.md) — the prompting walkthrough behind the [Single-shot in-between](#single-shot-in-between-two-key-poses) recipe.