ChatSpatial

An MCP server that drives end-to-end spatial transcriptomics analysis through natural language, wrapping Scanpy, Squidpy, and 60+ spatial methods.


Type	MCP server
Supplier	cafferychen777/ChatSpatial
Availability	GA — v1.2.10
Pricing	Free / OSS (MIT)
Capabilities	Read/Write — reads/writes `.h5ad` AnnData files on local disk and produces figures/result tables
Verified	works · 2026-07-20
Security	cleared · 2026-07-20 — provenance matches cafferychen777, PyPI chatspatial 1.2.10, MIT, no advisories

How to install

ChatSpatial runs as a local stdio MCP server. Install the PyPI package into a virtualenv first, then register it.

Claude Code — pip-based:
```
python -m venv venv
source venv/bin/activate
pip install chatspatial
which python   # note the absolute path printed
claude mcp add chatspatial /path/to/venv/bin/python -- -m chatspatial server
```
(replace /path/to/venv/bin/python with the absolute path that which python printed — e.g. /Users/you/ChatSpatial/venv/bin/python, or $(pwd)/venv/bin/python if you are still in the project dir.)
Claude Desktop — pip-based, in claude_desktop_config.json:
```
{
  "mcpServers": {
    "chatspatial": {
      "command": "/path/to/venv/bin/python",
      "args": ["-m", "chatspatial", "server"]
    }
  }
}
```
Fully quit and relaunch Claude Desktop after editing the config (it is not hot-reloaded). The pip install chatspatial step above is still required before this works.

Claude Desktop / Claude Code — Docker (no local Python env):

{
  "mcpServers": {
    "chatspatial": {
      "command": "docker",
      "args": [
        "run", "--rm", "-i",
        "-v", "/absolute/path/to/your/data:/data:ro",
        "-v", "/absolute/path/to/outputs:/outputs",
        "ghcr.io/cafferychen777/chatspatial:v1.2.10",
        "server", "--transport", "stdio"
      ]
    }
  }
}

Pull the image first with docker pull ghcr.io/cafferychen777/chatspatial:v1.2.10. Reference container paths (e.g. /data/sample.h5ad) in prompts, and use --rm -i (not -it) for MCP stdio. Replace the two /absolute/path/to/… mounts with your real input/output directories.

This is a long-lived stdio server launched by Claude itself — you do not run chatspatial server in a separate terminal; the registration above starts it on demand.

What it does

Exposes 20 schema-validated tools orchestrating ~65 spatial transcriptomics methods across 15 analytical categories: data loading and preprocessing, visualization, spatial domain identification, deconvolution, cell-cell communication, cell-type annotation, differential expression, trajectory inference, RNA velocity, spatial statistics, enrichment analysis, spatially variable gene (SVG) detection, multi-sample integration, CNV analysis, and spatial registration. Operates on AnnData (.h5ad) inputs.

Primary use cases: spatial transcriptomics preprocessing and QC, spatial domain identification, cell-cell communication analysis, spatially variable gene detection.

Notes

No API key required. Built on the scverse stack (Scanpy/Squidpy) plus additional spatial methods. Works with any MCP-compatible client (Claude Desktop, Claude Code, Codex). For non-spatial single-cell workflows see scanpy.md and the single-cell-rna-qc / scvi-tools plugins.

Sources

Installed this tool?

Share feedback — install path, OS, errors, workarounds. The form opens with this tool pre-selected and a link back to this page.