On this page

All notable changes to the Uthana API will be documented in this file.

## July 9, 2026

**Changed**

- `text-to-motion-3.0` is now available to any account on the pay-as-you-go plan. Use the new `create_text_to_motion_job` mutation with `model: "text-to-motion-3.0"`. See [Pricing](/content/docs/api/pricing/index.html) for details.P

## June 15, 2026

**Added**

- `text-to-motion-3.0` is now available in limited release via the new `create_text_to_motion_job` mutation. This async, diffusion-based model returns a `Job` that must be polled to completion, similar to video-to-motion. [Contact support](mailto:support@uthana.com) to request access.

**Changed**

- `video-to-motion-2.0` is now the default VTM model (previously `video-to-motion-v2`). `video-to-motion-v2` remains accepted as a backward-compatible alias.
- Added `video-to-motion-2.1` model with post-processing refinements. Returns refined motion IDs rather than the base motion.

**Removed**

- `video-to-motion-v1` is no longer available. The deprecation deadline of 30 Apr 2026 has passed.

## May 2026

**Added**

- `playback_speed` download option — scale animation playback speed between 0.1–2.0x (default 1.0).
- `rerig_target` and `include_fingers` options for character upload, for re-rigging to a specific target skeleton and including finger joints. Available in `@uthana/client` v0.3.1 and `uthana` (Python) v0.4.0.
- Client SDK usage examples added throughout the API docs (Python and JS/TS tabs).

## April 2026

**Added**

- `@uthana/client` and `@uthana/react` JS/TS client packages published on npm — v0.2.0 initial release.
- Python `uthana` package v0.2.0 — expanded API covering all capabilities with both async and sync variants.
- Locomotion API: `create_locomotion` mutation for controllable, loopable directional travel motions. JS/Python clients updated to v0.3.0 with locomotion support.
- `no_mesh` option for GLB downloads — returns animation data only, without the character mesh.

**Changed**

- `video-to-motion-v2` is now the default VTM model. `video-to-motion-v1` is deprecated and will be removed.

## February 2026

**Added**

- Stitch & loop motions: new `create_enhanced_stitched_motion` mutation transitions smoothly between two existing motions. See [Stitch & loop motions](/content/docs/api/capabilities/stitch-loop-motions/index.html).
- BUCMD `diffusion-v2` model for `create_text_to_motion` with advanced parameters: `steps`, `cfg_scale`, `length`, and `seed`.
- `auto_rig_confidence` field in character upload responses — a 0–1 quality score for the auto-rigging result.
- `in_place`, `torso_only`, and `roblox_compatible` download options for FBX/GLB exports.

## November 2025

**Added**

- Motion-only animation download endpoint: `/motion/animation/{character_id}/{motion_id}/glb/...` returns just the animation track without a character mesh, and does not count against download quota.

## January 30, 2025

**Added**

- Initial API documentation release
- GraphQL API reference
- Basic examples for cURL, Python, and TypeScript

**Changed**

- Documentation migrated to objective-oriented structure