Text to animation
Text to 3D animation: describe the motion, get the clip
Type what the character should do, such as "a tired run, then a skid to a stop", and Nira animates your own character, human or creature. Then refine it on the timeline and export.
On your character
Motion is generated for the rig you uploaded, not a stock mannequin you have to retarget later.
Constraints that stick
Pin a foot, hold a pose, or lock a contact, and the generated motion respects it.
Editable output
Every clip lands on a keyframe timeline you can adjust, loop, and sequence.
How text-to-animation works in Nira
For humanoid characters, Nira runs a text-to-motion model on your prompt and maps the result onto your rig. For quadrupeds, birds, and other creatures, Nira writes the animation as keyframes on the rig itself. Either way, you describe the motion and get an editable clip. Each generated clip runs up to 12 seconds; for longer performances, arrange several clips in the sequencer.
You talk to Nira the way you would brief an animator: "wave with the left hand and hold for two seconds", or "a looping idle for game export". Nira turns the brief into generation settings and puts the result on the timeline.
Direct the motion with constraints
Generated motion is only useful if you can control it. On humanoid characters you can lock a pose at a given frame, keep a foot planted, or pin a hand to a contact point, and the generation honours those constraints. That is how you get a character to land exactly on a mark or keep its feet from sliding.
What you can ask for
- Locomotion: walks, runs, sneaks, and turns, or a gallop, trot, or hop for creatures.
- Actions: jumps, punches, kicks, climbs, falls, pounces, and take-offs.
- Gestures and idles: waves, shrugs, tail flicks, and breathing loops.
- Blocking: quick first passes for previs and animatics.
Good to know
Motion constraints such as planted feet and pose locks apply to humanoid characters. Clips are capped at 12 seconds per prompt, and facial animation is not generated.
Frequently asked questions
How long can a generated clip be?
Up to 12 seconds per prompt. Chain clips in the sequencer to build longer performances.
How many tokens does a motion cost?
A motion costs 10 tokens plus 2 per second of animation, so a 3-second clip is 16 tokens and a 12-second clip is 34.
Can I use generated motion commercially?
You keep the rights to what you create in Nira. See the terms of service for the details that apply to your plan.
Does it work on non-humanoid characters?
Yes. Describe the motion and Nira animates quadrupeds, birds, and other creatures too, by writing keyframes on the rig instead of using its human motion model.
Bring your character to life
Upload a mesh and Nira rigs it, animates it, and exports it. New accounts start with 50 free tokens.