ChatSpatial Documentation

Agentic Workflow Orchestration for Spatial Transcriptomics Analysis

Get started now View it on GitHub

Key Features

  • Natural Language Interface: Analyze spatial data using conversational commands with Claude and other LLMs
  • Comprehensive Analysis Tools: Cell type annotation, spatial domains, trajectory analysis, and advanced spatial statistics
  • Multiple Data Formats: Support for H5AD, 10X Visium, Slide-seq, MERFISH, and other spatial transcriptomics formats

This documentation is currently being migrated to Just the Docs. Some links may be broken during the transition.

What is ChatSpatial?

ChatSpatial is an agentic workflow orchestration platform that eliminates the “implementation tax” in spatial transcriptomics research. Built on the Model Context Protocol (MCP), it integrates 60 state-of-the-art methods from fragmented Python and R ecosystems into a unified conversational interface. Unlike autonomous agents, ChatSpatial prioritizes human-steered discovery through schema-enforced tool-calling that ensures reproducibility and analytical fidelity.

Advanced Capabilities

  • Advanced Algorithms: Integration with state-of-the-art spatial analysis methods
  • Visualization: Built-in plotting and visualization capabilities

Supported Analysis Types

  • Cell Type Annotation: Marker-based, reference-based (Tangram, scANVI), and ML approaches
  • Spatial Domain Identification: SpaGCN, STAGATE, Leiden/Louvain clustering
  • Cell Communication: LIANA, CellPhoneDB, and spatial interaction analysis
  • Trajectory Analysis: RNA velocity, pseudotime, and developmental trajectories
  • Spatial Statistics: Moran’s I, spatial autocorrelation, and neighborhood analysis
  • Deconvolution: Cell2location, RCTD, Stereoscope for cell type proportions

Getting Started

Prerequisites

  • Python 3.10 or higher (required for MCP)
  • Claude Desktop or compatible MCP client
  • Git for installation

Quick Installation

# Clone the repository
git clone https://github.com/cafferychen777/ChatSpatial.git
cd ChatSpatial

# Create virtual environment (strongly recommended)
python3 -m venv chatspatial_env
source chatspatial_env/bin/activate  # macOS/Linux

# Install dependencies (recommended: full installation)
pip install -e ".[full]"

# Run the MCP server
python -m chatspatial

For detailed installation instructions, see the Installation Guide.

Your First Analysis

Once installed, you can start analyzing spatial data immediately:

Load the mouse brain Visium dataset and show me a spatial plot of the top variable genes

Identify spatial domains in my data using spagcn method and visualize the results

Perform cell type annotation using marker genes and show the spatial distribution

Documentation Structure

Getting Started

Tutorials

Reference

Examples

Community and Support

License

ChatSpatial is distributed under the MIT License. See LICENSE for more information.

About This Documentation

This documentation is built with Just the Docs, a documentation theme for Jekyll.