Skip to content

v0.0.8 — Clad body measurement, FLUX 3 Klein and Video, lip-sync, and 3D

Latest
Latest

Choose a tag to compare

View all tags
@kailashahirwar kailashahirwar released this 19 Oct 07:39
· 23 commits to main since this release
v0.0.8
ebcf4bf

OpenTryOn v0.0.8 Release Notes

Body measurement, FLUX 3 Klein and Video, lip-sync, 3D, and a larger understand / decide / embed set

Release Date: 19 October 2026

OpenTryOn v0.0.8 adds 13 model ids (15 registry slots, because FLUX 3 Image Klein is on both generate and edit) and three service categories: measure, audio-to-video, and generate-3d. 163 models, 168 MCP tools (163 models + 5 discovery / key / planner), up from 148 models / 153 tools. Thirteen services: vton, generate, edit, understand, video-generate, audio-to-video, generate-3d, bg-remove, tts, decide, embed, classify, measure.

What's New

New measure service — Clad body sizing

CLI model Path What it does
clad-body-photo Clad API (CLAD_API_KEY) One photo + height → async job → ISO 8559-1 measurements
clad-body-questionnaire Clad API Eight-question sizing flow, no photo
clad-body-local clad-body library Measure existing Anny or MHR param JSON. Not photo reconstruction

Image and video

CLI model Service What it does
flux3-image-klein generate + edit Distilled FLUX 3 Image, first-party BFL (BFL_API_KEY)
flux3-image-klein-local generate + edit Same family via Diffusers ModularPipeline
flux3-video video-generate FLUX 3 Video: text, keyframes, video-to-video, draft
sana-sprint generate Local NVLabs SANA-Sprint, 1–4 steps

New audio-to-video and generate-3d services

CLI model Service What it does
nvidia-lipsync audio-to-video NVIDIA Maxine LipSync NIM (self-hosted gRPC; no hosted REST)
trellis generate-3d Microsoft TRELLIS image/text → mesh (GLB/STL)

Understand, decide, embed

CLI model Service What it does
nemotron-3.5-super-vl understand NVIDIA NIM VLM (image, video, text; no audio). Same NVIDIA_API_KEY
videojev-4b, videojev-35b-a3b decide Local video grounding and multiple-choice QA. No hosted API
gemini-embedding-2 embed Hosted multimodal embeddings (text, image, audio, video, PDF). GEMINI_API_KEY

The docs site was also restyled to the TryOn brand system.

Breaking changes

  • No removed model ids and no new required dependency in install_requires.
  • New services are opt-in: measure needs CLAD_API_KEY for the hosted models; nvidia-lipsync needs a local NIM gRPC target; clad-body-local needs pip install 'clad-body[anny]' or 'clad-body[mhr]'.

Install / Upgrade

pip install -U opentryon
pip install -U "opentryon[local]"

From source:

git checkout v0.0.8
pip install -e .

Quick examples

opentryon measure --model clad-body-photo --photo full_body.jpg --height-cm 172
opentryon measure --model clad-body-questionnaire --age 30 --gender female --height-cm 168
opentryon generate --model flux3-image-klein --prompt "linen shirt, studio light"
opentryon video-generate --model flux3-video --prompt "model turning, soft daylight"
opentryon understand --model nemotron-3.5-super-vl --image look.jpg --prompt "Describe the outfit"
opentryon embed --model gemini-embedding-2 --text "black leather ankle boots"
opentryon generate-3d --model trellis --image product.jpg

Docs

Notes

  • Restart MCP after upgrade so clients see the new tools.
  • TryOn Studio capability screens still cover Image, VTON, Understand, Video, BG Remove, and TTS. measure, decide, embed, classify, audio-to-video, and generate-3d are available from the CLI and any MCP client, including Agent chat when you name the model.
  • 163 models across 13 services → 168 MCP tools.

Links

Assets 4
Loading