KonfAI documentation

KonfAI documentation

Run a model.
Build a workflow.

Medical-image segmentation, synthesis and registration. Start with a published app or build a reproducible workflow around your own model.

Run your first app → Installation

CT input Segmentation

Real data, real outputs. Explore the gallery ↗

Choose your task

• Published models

  Run a trained model

  Segment an image or generate a synthetic CT with a ready-to-use app.

  Run your first app →
• First experiment

  Train, predict, evaluate

  Complete a small CPU experiment with generated data and verified outputs.

  Follow the quickstart →
• Your own model

  Bring an existing model

  Use a PyTorch or MONAI model, or transfer compatible pretrained weights.

  Adopt KonfAI →
• Data preparation

  Prepare a dataset

  Resample, transform or combine volumes before your next experiment.

  Prepare your data →
• Image registration

  Align two images

  Choose an IMPACT-Reg preset and register a moving image to a fixed image.

  Use IMPACT-Reg →
• Interactive workflows

  Work through a conversation

  Inspect data, run jobs and view their results with Studio and the MCP server.

  Open the Studio guide →

Already using an MCP client? Connect the MCP server.

Keep these nearby

Settings, examples and answers

YAML configuration CLI reference Python API Example catalogue Troubleshooting Glossary