Text To Motion Diffusion v2 generates 3D character animation data from a plain text prompt. If you have spent hours manually keyframing basic movements in a 3D application, this model shortens that entire process to a single written description. Type what you want the character to do, select a figure from the built-in roster, and receive a ready-to-use animation clip. The model exports standard FBX or GLB files at 24, 30, or 60 frames per second, so the output slots into most 3D production and game development pipelines without conversion. Built-in inverse kinematics processing keeps feet grounded and movement physically plausible even on complex sequences. You can also dial in the cfg scale to make the animation follow your prompt description more or less strictly, giving you real control over the result. The output works well for game character rigs, cinematic previsualization, virtual avatars, and motion prototyping. Import the file into your software of choice, apply it to your character mesh, and iterate from there. Run the model as many times as you need until the motion feels right.
Text To Motion Diffusion v2 turns a plain text description into a fully rigged 3D character animation, ready to drop into your game engine or 3D pipeline. If you've ever spent hours keyframing a walk cycle or a combat idle, this model cuts that process to seconds. You type what you want the character to do, pick your settings, and get back an FBX or GLB file with the motion baked in. On Picasso IA, there's no software to install and no scripting required. It's the fastest path from a written description to an animation file your team can actually use.
Do I need programming skills or technical knowledge to use this? No, just open Text To Motion Diffusion v2 on Picasso IA, adjust the settings you want, and hit generate.
Is it free to try? Yes, you can run the model on Picasso IA without any upfront commitment. Check the current credit policy on the platform for details on how many free runs are available.
What output formats does the model support? The model exports in FBX and GLB format. FBX works well with most game engines and 3D applications, while GLB is the standard for web-based 3D and real-time rendering pipelines.
How long does it take to get results? Most generations complete in under a minute, depending on the motion length and server load at the time you run it. A 5-second animation at 30fps typically processes faster than longer sequences.
Can I control how closely the animation follows my prompt? Yes. The guidance scale lets you dial up or down how literally the model interprets your description. Higher values stick closer to the prompt; lower values give the model more freedom to produce fluid, natural motion.
What happens if the feet look like they're sliding? Enable foot inverse kinematics in the settings before generating. This anchors the feet to the ground plane and removes most sliding artifacts from the output.
Where can I use the animations I generate? FBX and GLB files are compatible with most 3D tools and game engines. You can import them into your animation pipeline, use them in a game project, or incorporate them into a cinematic sequence.
Everything this model can do for you
Type any motion description in plain language and receive a full 3D character animation file.
Export files in two industry-standard formats that load directly into 3D applications and game engines.
Choose 24, 30, or 60 fps to match the target playback standard of your project.
Pick from Tar, Ava, Manny, Quinn, or Y Bot to suit different character styles and output needs.
Inverse kinematics is applied automatically so feet stay grounded and motion looks physically natural.
Set clip duration to fit loops, single actions, or longer cutscene sequences.
Increase cfg scale to make the output follow your written description with greater precision.
