How is text to motion different from motion capture?

Text to motion synthesizes a plausible new performance from words. Motion capture reconstructs a specific recorded performance from video or sensors.

Text to Motion

Q: What is text to motion?

Text to motion is generative animation: an AI model interprets a written action and produces a time-based 3D skeletal movement sequence.

Q: Does generated motion need cleanup?

Often yes. Review balance, root movement, contacts, retargeted pose, loops, props, terrain and target-character deformation before production use.

Describe a human action in plain language and turn it into editable 3D skeletal motion. Generate walk cycles, gestures, sports moves, combat actions, dance ideas and robot behaviors without filming a performer.

Try Now →

What Is Text to Motion?

Text to motion is generative animation that translates a written description—such as “walk forward confidently, stop, turn left, then wave”—into one or more plausible 3D skeletal-motion sequences. It creates a new performance rather than reconstructing a recorded actor.

QuickMagic's open-beta workflow lets you refine a prompt, choose duration and candidate quantity, compare generated takes, select an export preset and download the chosen source motion. Retargeting and contact cleanup remain separate production steps.

Status: Open beta announced June 5, 2026
Free use: 15 seconds per user per day
Duration: Up to 10 seconds per generated animation
Candidates: Up to 10 alternatives per project
Prompt tools: Refine/Polish, Sample Script and multilingual input

How Text to Motion Works

Move from one clear human-action prompt to an exportable motion candidate in three steps.

Step 1

Writing and refining a text-to-motion prompt

Write or Refine

Start with “A person,” use concrete verbs, join short sequences with “and then,” and add direction, speed, posture or an ending pose. Use Refine/Polish or Sample Script when needed.

Step 2

Choosing duration, candidate count and an export preset

Configure and Generate

Choose a duration that fits the action, set the number of alternatives and select the closest target skeleton or export preset before generating.

Step 3

Comparing, exporting and retargeting AI-generated skeletal motion

Compare and Export

Check prompt alignment, balance, root trajectory and contacts. Export the best source motion, preserve an untouched copy, then retarget and refine it.

Try Now →

From Action Prompt to Skeletal Motion

Write observable movement rather than appearance. A strong prompt names who moves, what happens in sequence, how fast it happens and where the motion finishes.

PROMPT ↓

A person steps back with the left foot, blocks high with both arms, and then pivots forward into one controlled right-leg kick before returning to a balanced guard.

Weak prompt	Why it is unclear	Stronger motion prompt
A heroic fight.	No physical action, side or sequence.	A person steps forward, blocks with the left forearm, throws a right cross, and returns to guard.
Walk nervously.	The emotion does not define body mechanics.	A person walks forward with short hesitant steps, shoulders raised, and pauses to look over the left shoulder.
Cool dance.	Style, timing and body parts are undefined.	A person steps right, rotates the torso, swings both arms outward, and finishes in a wide low pose.

Try Now →

Select the Best Motion Candidate

Variation is expected: one prompt can describe several valid performances. Generate alternatives before spending time polishing one clip.

Candidate Review Checklist

Action order, body side, direction, speed and ending pose match the prompt.
Balance and weight look plausible through steps, bends, jumps and impacts.
Root travel and rotation fit the scene and target character.
Feet, hands and imagined objects contact at usable frames.
Start and end poses can blend into neighboring clips.

Export and Production Cleanup

Preserve an untouched source-motion export.
Import with the correct frame rate, axis and unit scale.
Match source and target T-pose or A-pose before retargeting.
Correct stride, root movement, loops, foot sliding and prop contacts.
Reimport the final animation into a clean scene for validation.

Generate Motion Before You Animate

Explore Actions Without a Camera

Turn written choreography into motion candidates when an actor, studio or suitable reference clip is unavailable. Test different pacing and performance ideas before committing to manual animation.

Text prompt interface used to generate humanoid motion

Build a Character Motion Library

Create drafts for locomotion, idles, gestures, dance, sports and combat. Retarget the chosen skeletal motion to compatible game characters, digital humans, MMD models or VTuber avatars.

Generated motions applied to different humanoid characters

Compare Multiple Interpretations

Language can describe several valid performances. Generate alternatives, compare action order, direction, timing, balance and root movement, then select the most useful candidate.

Multiple AI-generated 3D motion candidates for review

Continue in Your 3D Pipeline

Export animation for compatible Blender, Unreal Engine, Unity, Maya, Mixamo, MMD, iClone and humanoid-robot workflows. Retargeting and final contact cleanup may still be required.

Text-generated skeletal motion exported to a 3D production pipeline

Try Now →

Text to Motion Examples

Explore generated action drafts and downstream character, MMD and humanoid-motion workflows.

Try Now →

Text to Motion FAQ

What is Text to Motion?

It is an AI motion-generation workflow that turns a natural-language action description into a time-based humanoid skeleton sequence for preview, retargeting and editing.

How is it different from video motion capture?

Video mocap reconstructs a specific performance visible in footage. Text to motion synthesizes a plausible new performance, so the same prompt can produce multiple valid results.

What makes a good motion prompt?

Name the subject, sequence of actions, direction, speed, weight, mood and ending pose. Avoid relying on visual appearance because the output is motion data.

Is QuickMagic Text to Motion free?

The June 5, 2026 launch information lists a free allowance of 15 generated seconds per user per day. Current product limits and plans are the final source of truth.

How long can one generated motion be?

The open-beta launch interface supports up to 10 seconds for one animation and up to 10 candidate animations per project.

Which languages can I use?

QuickMagic describes multilingual prompt input. Refine/Polish can rewrite an instruction into a clearer motion-oriented format, although results may vary by language.

Does Text to Motion generate a 3D character?

No. It generates skeletal motion data or animation for an export preset. Character modeling, rigging, skinning, materials and rendering are separate.

Can it create exact choreography?

It is best for motion ideas and drafts. Use video mocap or manual keyframing when exact performance, trajectory, timing or contact must be reproduced.

Which export workflows are supported?

The launch interface shows presets including FBX characters, Mixamo, Unreal and BIP. The broader QuickMagic product supports additional workflow-specific formats; availability varies.

Does generated motion need cleanup?

Often yes. Review balance, root movement, foot and hand contacts, retargeted pose, loops, props, terrain and target-character deformation before production use.

Try Now →