Generated Computers
Locals: an Apple Silicon engine that runs local LLMs up to 3x faster. A product of Generated Computers.
Translation pending
OpenJaws: An Open Hardware UMI Framework with Stereo Localization for Generalist Manipulation
Developer toolsJavaScript · HTMLGitHub @open-jawsOnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Run /nomorevibe verify in your project folder to verify ownership and manage the introduction yourself.
Verify in one minuteAsk about other verification options
Website for OpenJaws
Translation pending
Taken from the repository README
Loading the full README.
Read the full README ↗Project website for OpenJaws: An Open Hardware UMI Framework with Stereo Localization for Generalist Manipulation.
Static site, no build step — index.html plus static/. Served directly by GitHub Pages.
python3 -m http.server 8000
# then open http://localhost:8000
| Section | Contents |
|---|---|
hero (#top) | Title, authors (all linked), NVIDIA logo, link buttons |
#teaser | UMI-Loc-Bench-v1 clip, autoplaying as the opening visual |
#abstract | System-overview figure, then the abstract and a four-card stat strip |
#video | Full 55 s video with chapter chips that seek the player |
#data-collection | Chapter 1 clip |
#system | Three design-principle cards + gripper mechanism |
#localization | Stereo triangulation + stereo-capture figure |
#benchmark | UMI-Loc-Bench-v1 prose + task-sequence figure |
#results | Translational-ATE plot + stereo/monocular/neural discussion |
#real-robot | Chapter 3 clip + deployment summary |
#real2sim | Chapter 4 clip + digital-twin summary |
#bibtex | Citation |
static/videos/openjaws_overview.mp4 is the full 55 s supplementary video, cut from
../ICRA27_8587_VI_i.mp4 at the chapter boundaries below. The source is silent (no audio track);
chapters were recovered from its on-screen title cards.
| Clip | Range | Chapter | Used in |
|---|---|---|---|
chapter_1_data_collection.mp4 | 0:04–0:13 | Robot-free data collection | #data-collection |
chapter_2_benchmark.mp4 | 0:13–0:29 | UMI SLAM benchmark | #teaser |
chapter_3_real_robot.mp4 | 0:29–0:45 | Real robot experiments | #real-robot |
chapter_4_real2sim.mp4 | 0:45–0:55 | Real-to-sim comparison | #real2sim |
The four chapter clips play GIF-style: autoplay muted loop playsinline, no controls. They are
MP4 rather than real .gif because a 16 s 720p GIF runs 40–80 MB against ~4 MB for the MP4.
Only the full video in #video keeps controls; its chapter chips seek via the bounds array in
the inline script at the bottom of index.html. If the source video is ever re-cut, re-export the
clips, re-grab the static/images/poster_*.jpg frames, and update that array.
The hero link buttons are still placeholders (href="#") — fill in arXiv, paper PDF, code,
hardware/BOM, and dataset URLs. The BibTeX entry needs its real arXiv ID in place of
arXiv:XXXX.XXXXX.
The author list uses the camera-ready names from ../paper/main.tex; swap it for
"Anonymous Author(s)" if the site must go up while the submission is still anonymous.
Content is sourced from ../paper/. See ../website_plan.md for the build plan and the figure
and video pipeline.