Uthana
Uthana is an AI platform built on foundation models for human motion. It turns text prompts, 2D video or live controller input into 3D animation, then auto-rigs and retargets the result onto any biped character.
What is Uthana?
Uthana, Inc. runs an AI platform devoted to a single problem: making 3D characters move. Its foundation models for human motion are trained on large-scale motion capture and can be conditioned on text, 2D video, voice or game state, so a written description, a phone clip or a gamepad becomes animation data rather than a rendered video.
The suite is organised around eight products. Text to motion turns a prompt into animation and returns three variants from different models so you can choose. Video to motion reconstructs 3D movement from a single monocular clip. Locomotion and real-time animation let you drive a character with mouse, keyboard or gamepad and record the take. Stitching blends clips, looping builds seamless cycles, auto-rigging generates a skeleton for an unrigged mesh, and flexible retargeting maps any motion onto any biped rig through a proprietary IK system.
Everything runs in the browser, with no plugin or install required, and a 3D viewer for previewing, trimming, looping and rotating bones before export. Files come out as FBX or GLB, with BVH on request, keyframes intact for Maya, Blender, Unreal, Unity or Roblox. Developers get a GraphQL API, official Python and TypeScript clients, React hooks, and open-source Blender and Maya plugins; SDKs for Unreal and Unity are announced as being in private beta.
Underneath sits a dataset Uthana says it owns outright: more than 500 hours of studio mocap, clips at 60 to 120 FPS running 0.5 to 45 seconds, skeletons of 22 to 65 joints, cleaned, labelled and physics-annotated in house and sourced with performer consent. The company claims real-time inference under 15 ms, most text-to-motion generations finishing in under ten seconds, over a million seconds of motion generated and more than 15,000 characters rigged and animated.
Three markets are addressed explicitly: game and animation studios, AI-native creation platforms that call Uthana as a motion layer, and robotics or embodied-AI teams exporting to CSV, URDF or MJCF for simulation and imitation learning. Uthana is candid about the boundary: its models cover realistic human movement on biped humanoids, and it describes itself as a creative multiplier, not a replacement for animators.
What it does
- Generate 3D human motion from a plain text prompt
- Extract 3D motion from a single 2D video clip
- Auto-rig an unrigged biped mesh in seconds
- Retarget any motion onto any biped skeleton
- Stitch, loop and trim motions with keyframe control
- Drive and record a character live with mouse, keyboard or gamepad
- Export production-ready FBX, GLB or BVH files, or call everything through a GraphQL API
When to use Uthana / When not to
A quick filter to help you decide if Uthana is the right fit.
When to use Uthana
- Indie game developers who need varied character animation without a motion-capture budget
- Animation studios wanting to cut keyframing and mocap cleanup out of an existing pipeline
- 3D animators and rigging artists dealing with unrigged or non-standard biped characters
- Engineering teams at AI-native creation platforms adding a motion layer through an API
- Robotics and embodied-AI researchers who need large volumes of retargetable humanoid motion
When not to use Uthana
- Anyone animating quadrupeds, creatures or non-humanoid rigs, since the models only handle biped humanoid skeletons
- Creators chasing stylised, fantastical or physically impossible movement, which the vendor admits it handles poorly
- Free-plan users planning a commercial release: outputs generated on the free tier belong to Uthana and ship under CC BY 4.0
- Organisations bound by the GDPR, since the site never mentions it, offers no DPA and declares no hosting jurisdiction
- Anyone under 18, who is barred from the platform by the terms of use
How to use Uthana
A typical end-to-end flow, from setup to results.
- Create a free Uthana account
- Upload a character as .glb or .fbx, in T-pose or A-pose with feet on the ground, or start from a stock character such as Tar, Ava, Manny or Quinn
- Let auto-rigging build a skeleton if the mesh has none, which takes roughly 30 to 60 seconds
- Pick a capability: a text prompt, a 2D reference clip, a stitch between two motions, or live control with keyboard or gamepad
- For text to motion, open the advanced options to set foot contact (IK), prompt enhancement, motion length and diffusion steps
- Compare the three variants the web app returns and keep the one that fits
- Preview in the browser viewer, then trim, loop, adjust speed or rotate bones
- Add the motions worth keeping to your library
- Export as .glb or .fbx, BVH on request, and import into your tool or engine
- To automate, generate an API key in your account settings, install the Python or TypeScript client and call the GraphQL endpoint, or work through the Blender and Maya plugins
Pros & Cons
Pros
- A genuinely usable permanent free plan: unlimited generation, 30 download seconds a month, 10 uploads or auto-rigs, and API access
- Fast: most text-to-motion runs finish in under ten seconds, and real-time inference is quoted under 15 ms
- Nothing to install, since the web app and 3D viewer run entirely in the browser
- Standard, editable outputs in FBX and GLB that drop straight into Maya, Blender, Unreal, Unity or Roblox
- Auto-rigging removes the step that usually blocks AI-generated characters
- A documented GraphQL API with official Python and TypeScript clients and open-source plugins
- Usage-based pricing published product by product, alongside the subscription tiers
Cons
- Free-plan outputs belong to Uthana and only come back under a CC BY 4.0 licence, with attribution required
- Customer data feeds model training by default; the no-training tier is mentioned conditionally and no opt-out is documented
- No mention of the GDPR anywhere, no data processing agreement, no EU representative and no formal sub-processor list
- No hosting jurisdiction is declared, so there is no way to know where uploaded characters and videos live
- No postal address, no legal notice and no contact page: support and cancellation run through email alone
- The pricing page contradicts itself, with struck-through prices, duplicated tiers and a parallel set of plans marked Coming soon
- Limited to realistic human movement on biped humanoid rigs, with a 100 US dollar monthly minimum on pay-as-you-go
Pricing & Plans
Uthana offers a permanent free plan, Dreamer, rather than a time-limited free trial: the pricing FAQ answers the trial question by pointing to it. The cheapest paid entry point is the Indie tier at USD 9.00 per month, displayed against a struck-through USD 15.00, or USD 72.00 billed annually. A pay-as-you-go alternative is billed per generated second and carries a USD 100 minimum monthly commitment. All prices are published in United States dollars.
- unlimited motion generation
- 30 download seconds a month
- 10 character uploads or auto-rigs
- 2 character generations
- API and SDK access
- non-commercial Creative Commons licence
- Discord support
- 50 to 75 download seconds a month and 40 to 50 auto-rigs depending on the table
- Creative Commons or commercial licence depending on the block
- email and Discord support
- everything in Indie plus 150 download seconds a month
- 125 character uploads or auto-rigs
- commercial licence
- everything in Creator plus 500 download seconds a month
- 400 character uploads or auto-rigs
- priority support
- everything in Pro plus 2
- 000 download seconds a month
- 1
- 500 character uploads or auto-rigs
- 12
- 500 download seconds a month
- 7
- 500 uploads or auto-rigs
- team accounts
- dedicated Slack channel with an SLA
- high-volume API and SDK integration
- data siloing
- text to motion USD 0.02 per second (v1.0)
- USD 0.03 (v2.0) and USD 0.10 (v3.0)
- video to motion USD 0.05 per second
- locomotion USD 0.02 per second
- stitch and loop USD 0.01 per second
- auto-rigging USD 0.15 per rig
- character generation USD 0.75
- retargeting free until October 2026
- a second set of cards on the same page lists Student at USD 10
- Starter at USD 16
- Pro at USD 50
- Builder at USD 160 and Platform at USD 660 a month
- all marked Coming soon with a waitlist
Data, GDPR & hosting
A consolidated view of how Uthana handles your data.
GDPR overview
There is no GDPR implementation to report. Across every page collected, from the privacy policy and terms of use to the FAQ, pricing and product pages, the words GDPR and General Data Protection Regulation never appear, and neither does a data processing agreement, a sub-processor list, an Article 27 EU representative or a data protection officer. The privacy policy, version 1.0 dated 12 November 2024, rests entirely on United States state privacy law: it carries dedicated California Resident Rights and Nevada Resident Rights sections and states that Do Not Track signals are not honoured. Uthana neither claims compliance nor denies it. The only privacy contact published is the general support address. European organisations should read this silence as unresolved rather than as a negative answer.
Who owns the data?
Uthana's terms split ownership three ways. You keep all right, title and interest in your Customer Inputs, the prompts, videos, mocap and 3D models you upload, subject to the licences you grant Uthana. Outputs generated on a paid plan are yours, with a perpetual worldwide licence over any Uthana intellectual property embedded in them. Outputs generated on the free plan belong to Uthana, which licenses them back under CC BY 4.0. Uthana retains everything in its platform, models, algorithms and training data. Payment details go to Stripe, Inc., Google LLC sets analytics cookies, and all personal data may pass to a third party in a merger, acquisition or bankruptcy.
Reuse rights
On a paid plan you may reuse the generated motion freely and commercially without asking permission: Uthana grants a non-exclusive, worldwide, perpetual, royalty-free licence over any of its intellectual property embedded in the output. On the free Dreamer plan the same motion comes back under a Creative Commons Attribution 4.0 licence, which the pricing page presents as non-commercial and which requires attribution. Files export as FBX or GLB, with BVH on request, and keyframes remain fully editable in your own tool. In the other direction, Uthana may use your Customer Data, personal information excluded and anonymised where applicable, as training data for its machine learning models, and no opt-out mechanism is documented.
Data retention & training
Hosting summary
Uthana declares no hosting jurisdiction. Neither the privacy policy nor the terms of use names a country, a region or a data centre, and the site has no trust or security page. The privacy policy lists only categories of service providers, covering hosting, technology and communications, analytics, security, support, fulfilment and payment, and names just two of them: Stripe, Inc. for payments and Google LLC for Analytics. No sub-processor list, no SOC 2 or ISO 27001 certification and no data processing agreement appears anywhere in the collected pages. Security is described in generic terms as physical, technical, organisational and administrative measures. Technical signals point to the United States without amounting to a commitment: the domain resolves to an address geolocated in Boardman, Oregon on Amazon's network, the registrar is Amazon Registrar, Inc., the marketing site is served by Webflow, and an AWS Activate badge sits in the homepage footer. Enterprise customers are offered data siloing and, on the data page, on-premise deployment, with no further detail. Anyone with a residency requirement should ask before uploading.
Things to keep in mind
Risks and trade-offs to weigh before adopting Uthana.
- Your uploads feed model training by default, so a studio can hand over a signature style or an unreleased character without meaning to
- Free-plan outputs belong to Uthana and require CC BY 4.0 attribution, which makes building a commercial project on them a licensing trap
- No GDPR framework, no data processing agreement and no sub-processor list, leaving anyone handling European personal data with nothing to rely on
- No declared hosting jurisdiction, so the location of uploaded characters and reference videos is simply unknown
- Uploading a reference video can expose a real person's likeness, and the site sets no consent rules for the people filmed
- Pipeline dependency on a young third-party API, with a 100 US dollar monthly minimum on pay-as-you-go and promotional pricing that expires, retargeting being free only until October 2026
- Skill erosion and sameness: instant generation can discourage learning keyframing and acting, and models trained on one mocap corpus tend towards average, interchangeable movement
Setup & Integrations
Technical difficulty
Very low for the web app: create an account, upload a .glb or .fbx character in T-pose or A-pose with its feet on the ground, and generate. Nothing is installed, and no rigging or animation skill is needed for a first result. Moderate for the API: you generate a key in your account settings, install the Python or TypeScript client and issue GraphQL calls with Basic authentication. Blender and Maya plugins install from GitHub. There is no official C# client, so engine-side work means raw GraphQL over HTTP.
Deployment
Integrations
Behind Uthana
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does Uthana actually produce?
Do I need a rigged character to use it?
Is there a free plan or a free trial?
Who owns the motion I generate?
Is my data used to train Uthana's models?
Is there an API?
How fast is generation?
What kinds of motion does it handle badly?
Is Uthana GDPR compliant?
Is there a minimum age?
Should you pick Uthana?
Uthana does one thing, and does it with unusual depth. Where most AI creation tools stop at a static character, it covers the whole motion chain: generation from text or video, auto-rigging, IK retargeting, in-browser editing and export in the formats a real pipeline expects. The engineering signals are credible, with a documented GraphQL API, official Python and TypeScript clients, open-source Blender and Maya plugins, an active blog and a Unity AI partnership announced in May 2026.
The commercial side is less settled. The pricing page stacks several tables at once, with struck-through prices, duplicated tiers and a parallel set of cards marked Coming soon, and the site contradicts itself on whether API access starts on the free plan or on Pro. The size of the motion library shifts from 100,000 to 75,000 to 50,000 clips across pages of the same site.
The real caution is governance. Customer data feeds model training by default, with no documented opt-out. Outputs generated on the free plan belong to Uthana and return under a CC BY 4.0 licence. The GDPR is never mentioned, no data processing agreement is offered, no sub-processor list is published, no hosting jurisdiction is declared, and the company publishes no postal address or legal notice while imposing arbitration in California.
For an indie developer or a studio wanting to compress days of keyframing into seconds, the free tier makes the decision cheap to test and the paid entry point is modest. For a regulated organisation, or for anyone whose uploaded material is sensitive, the data terms deserve a careful reading before the first upload.
- Choosing a selection results in a full page refresh.
- Opens in a new window.