wisplab
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@wisplabquantify my Western blot at ~/Downloads/wb.png with 6 lanes and rolling ball r=50"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
wisplab-cli
WispLab 三引擎的无头命令行 + MCP server。零运行时依赖:分发到各仓库自己的 tsx/vite/playwright,保证跑的永远是官方引擎代码。CLI 与 MCP server 共用同一份调度(lib.mjs)。
人不需要学本 README:四个 agent 技能(
~/.agents/skills/wisplab-{wb-quantification,ihc-quantification,if-colocalization,qpcr-analysis}/)已把「用户的话 → 正确命令 → 结果解读」全流程教给 AI——直接对 AI 说"帮我定量这张膜/这张切片/这张 qPCR 表"即可。本文件只是命令的参考手册。
安装(一次性)
git clone https://github.com/Gaoyuan-0423/wisplab-cli.git && cd wisplab-cli
cp wisplab.config.example.json wisplab.config.json # 改成你的三个仓库路径
# 三个引擎仓库就位并 npm install 后:
npm link # 可选;也可直接 node cli.mjs新人上手与自然语言用法见 [GETTING-STARTED.md](GETTING-STARTED.md)。
Related MCP server: Image Analysis MCP Server
命令与已校准用法
# WB:6 泳道 β-Actin 图,rolling ball r=50,泳道 1 为对照
wisplab wb CST-cst-4970-beta-actin-01.png --strip "140,85,160,350" --lanes 6 --bg rolling:50
wisplab wb CST-cst-4970-beta-actin-01.tif --strip "140,85,160,350" --lanes 6 --bg local # 与 M4 记录逐值一致(≤1e-05)
# qPCR:双内参 + 两对照样本(与 M1 独立核对记录逐值一致 0.00e+00)
wisplab pcr RELATIVE_example.csv --ref GAPDH,ACTB --control AiW002-2-D0,KYOU-D0
wisplab pcr 表.csv --ref GAPDH --eff "LDHA=200" # Pfaffl 效率校正
# IHC/染色:H-DAB 整图定量(与 M3 记录逐位一致);--control-tif 走对照片标定
wisplab ihc ptglab_ki67_110311.jpg
wisplab ihc 样本.tif --control-tif 无染色对照片.tif
# IF 双通道共定位(独立子命令;与 M3 CD20-CD3 记录同引擎)
wisplab if duplex.jpg--strip y,h,x0,x1:多面板截图先人工定位目标面板条带行(含上下背景余量),x 边界排除 marker 梯带与文字列,引擎在条带内精确分割泳道。单面板图可省略,让引擎全图提议。
MCP server(ZCode / Codex / WispScience 接入)
node ~/Downloads/WispLab预印本验证/wisplab-cli/mcp-server.mjs五个工具:wb_quantify/ihc_quantify/if_colocalize/pcr_analyze/pcr_verify。客户端配置示例(ZCode):
{ "mcpServers": { "wisplab": { "command": "node",
"args": ["<你的路径>/wisplab-cli/mcp-server.mjs"] } } }安全:文件路径必须在 allowPaths(config,默认 ~/Downloads 与 /tmp)内,域外路径直接拒绝。
校准状态(2026-10-18)
入口 | 基准 | 结果 |
| M4 记录 app_results.json | net 相对差 ≤1e-05 |
| M4 parity 语义 | 值低于 local 属预期;单步一致性见仓库 parity 记录 |
| M1 记录 app_output_S2 | FC 与 p 值逐值 0.00e+00 |
| M3 记录 results_m3 | 阳性面积/阈值逐位一致 |
| M3 记录 results_m3/if | 同引擎冒烟通过 |
MCP 握手/工具调用/路径防护 | /tmp/mcp-test.mjs | initialize+tools/list+call 全过 |
已知边界
ihc/if首次运行自动拉起 vite dev(约 2–15 秒),结束时自动回收;端口冲突改 configihc.devPort。語料 JPEG 输入对 wb 是设计内拒绝(仅 TIFF/PNG);ihc/if 接受 JPEG/TIFF/PNG。
输出均为 JSON(stdout);叠加图当前不导出(需要时用
scripts-ihc/run_m3.mjs)。
This server cannot be deployed
Maintenance
Related MCP Connectors
Connect AI clients to biomedical data and tools.
Image & PDF tools for AI agents: compress, convert, resize, PDF, AI vision, pipeline.
PDF, image, video, OCR, screenshot, SQL, QR and text tools for agents. No API key, no signup.
Use your Mac, Windows or Linux computer from ChatGPT, Claude or Codex: files, commands, documents.
Related MCP Servers
- AlicenseAqualityAmaintenanceEnables AI agents to control Fiji/ImageJ for microscopy image analysis through natural language commands, supporting operations like image opening, filtering, particle analysis, and automated workflows.1936 PyPI2BSD 3-Clause
- AlicenseNot gradedqualityDmaintenanceEnables AI assistants to locally process images with tools for cropping, zooming, enhancement, edge detection, segmentation, and text region extraction, all without external API keys. It uses PIL, OpenCV, and scikit-image for robust image analysis.MIT
- AlicenseAqualityCmaintenanceProvides AI agents with local file-processing capabilities for token counting, RAG chunking, CSV/JSON conversion, QR generation, and more, while keeping documents private on the user's machine.71MIT
- FlicenseNot gradedqualityCmaintenanceEnables in-chat scientific viewers for molecular structures, sequence alignments, and slide images, plus data tools for searching structures, fetching sequences, aligning them, and drafting NGS analysis plans.-