Repository navigation
Releases
LabscriptAI OT v0.1.3
One lightweight package with detailed Flex models. Install once; matching updates reuse dependencies and user records. Version and build are visible in health_check. Citation: Gao et al., bioRxiv 2025, doi:10.1101/2025.09.30.679666.
LabscriptAI OT v0.1.2
更省心的安装与更新,以及更清楚的 Flex 三维预览。
- 一次安装,后续复用:依赖和用户记录存放在稳定目录,更新插件不必重复下载相同依赖。
- 版本一目了然:安装检查和健康检查显示实际加载的版本与构建号,配置指向旧副本时给出提示。
- 三维画面更直观:白色背景、台面/整机视角、PE 枪头盒与开口枪头、EP 管和混合离心管架;修复 D3 未使用废物滑道造成的重叠。废液容器使用独立槽位。
- Windows 预览启动修复:正确识别 Python 启动器后面的实际服务进程,避免服务已经启动却误报超时。
- 结果更易读:完整 warnings 数组与数量、可指定时间区间的视频、清楚区分放回与丢弃枪头。
- 流程更灵活:按用户明确选择执行软件检查,保留真实停机信号和机器人状态提示。
下载下方 labscriptai-ot-0.1.2.zip,解压后在 labscriptai-ot 目录运行 install-labscriptai-ot.sh(macOS/Linux)或 install-labscriptai-ot.ps1(Windows)。已有三维预览请关闭后重新打开。
一个约 2.97 MB 的安装包已包含精细模块模型。包内不包含测试文件、临时数据、密钥和依赖环境。详细安装说明见 Getting Started。
本地验证:Node 402 项、Flex Python 41 项、打包 4 项通过,并通过独立复测。Windows、macOS、Linux CI 的全新发布包安装及本地完整流程均通过。本版本未连接真实机器人验证;耗材外观为程序化模型,不代表厂家 CAD。
Citation: Gao, Yuan et al. “Autonomous Liquid-handling Robotics Scripting for Accessible and Responsible Protein Engineering.” bioRxiv, 2025. doi:10.1101/2025.09.30.679666. Distributed under the LabscriptAI Research Citation License v1.0.
LabscriptAI OT v0.1.1
Download labscriptai-ot-0.1.1.zip — one 2.80 MiB bundle for Codex, Claude Code and Cursor, with all five detailed Flex module models included and loaded automatically. Extract it, then run the included installer; Node/Python dependencies install locally. No separate model pack is needed. Exact size and SHA-256 are in release-manifest.json.
- 52 MCP tools for authoring, local simulation, LogicPass, 3D checks and opt-in runtime assistance.
- Clearer questions about missing experiment facts; compact results distinguish local checks from physical readiness.
- 3D previews remain available after the launching client exits, with a private startup log for diagnosis. Browser scene loading no longer creates a server-side OpenGL renderer.
- Combines the branch/PR improvements for modules/adapters, waste chute, recovery guidance and monitoring.
Validation: 359 Node tests, 9 Python tests and 3 packaging tests pass. Fresh ZIP installation passed 19 setup checks; detailed meshes/motion passed five additional tests against the extracted package. Real DeepSeek Harness scenarios covered clarification, eight independent sample transfers with simulation/LogicPass/3D checks, minimal repair, missing live prerequisites and persistent preview. Browser rendering and step navigation were verified after the Harness exited.
No physical robot execution or full host-client UI installation is claimed as tested. Live robot actions remain opt-in; local checks do not establish assay accuracy or physical readiness.
License: LabscriptAI Research Citation License v1.0. Citation required: Gao, Yuan et al. Autonomous Liquid-handling Robotics Scripting for Accessible and Responsible Protein Engineering. bioRxiv (2025), https://doi.org/10.1101/2025.09.30.679666 .
labscriptai-ot v0.1.0
LabscriptAI OT bundles Opentrons MCP, seven agent skills, safety policy, and a curated protocol library.
Author: gaoyuan
Contact: gaoyuanbio@qq.com
License:
- This release is not MIT licensed.
- It uses the LabscriptAI Research Citation License v1.0.
- Public use, derivative work, publications, reports, benchmarks, demos, repositories, and AI tool/plugin integrations must cite the LabscriptAI bioRxiv preprint.
- DOI: 10.1101/2025.09.30.679666
Safety:
- Live robot execution is opt-in.
- Simulation remains the gate before unattended live execution.
- Vision is optional and observation-only.
Install:
- Use the GitHub source archive for this tag, or clone the repository.
- Run
bash install-labscriptai-ot.shfrom the repository root.
Platform metadata included:
- Cursor:
.cursor-plugin/,.cursor/ - Claude Code:
.claude-plugin/ - Codex:
.codex-plugin/,.mcp.json