ChatSpatial
README.md
<div align="center">
# ChatSpatial
**MCP server for spatial transcriptomics analysis via natural language**
[](https://doi.org/10.64898/2026.02.26.708361)
[](https://openreview.net/forum?id=xZ814yNaUW)
[](https://www.enar.org/meetings/spring2026/)
[](https://www.ibc2026.org/home)
[](https://github.com/cafferychen777/ChatSpatial/actions/workflows/ci.yml)
[](https://pypi.org/project/chatspatial/)
[](https://www.python.org/downloads/)
[](https://opensource.org/licenses/MIT)
[](https://docs.cafferyang.com/)
[](https://github.com/users/cafferychen777/packages/container/package/chatspatial)
</div>
<p align="center">
<img src="assets/images/overview.jpg" alt="ChatSpatial Overview" width="900">
</p>
ChatSpatial replaces ad-hoc LLM code generation with **schema-enforced orchestration**. Instead of generating arbitrary scripts, the LLM selects tools and parameters from a curated registry, making spatial transcriptomics workflows more reproducible across sessions and clients.
ChatSpatial exposes **20 schema-validated MCP tools** that orchestrate **67 spatial transcriptomics methods** across **13 analytical categories**. The tools are the stable natural-language interface; the methods are the analysis backends selected through tool parameters.
The server implements MCP `2026-07-28` through the official Python SDK v2 and
continues to serve `2025-11-25` clients through SDK-managed protocol negotiation.
STDIO remains the secure local default; Streamable HTTP is available for
explicitly configured HTTP deployments.
---
## Start Here
Install [`uv`](https://docs.astral.sh/uv/getting-started/installation/) once, then
register ChatSpatial without creating or managing a Python environment:
**Codex:**
```bash
codex mcp add chatspatial -- uvx --from chatspatial chatspatial server
```
**Claude Code:**
```bash
claude mcp add --scope user chatspatial -- \
uvx --from chatspatial chatspatial server
```
`uvx` creates an isolated environment on first launch and reuses its cache on
later launches. Restart the MCP client after adding the server.
The command above installs the standard runtime. To make all 15 composable
Python method families available in the same isolated MCP environment, use:
```bash
uvx --from 'chatspatial[full]' chatspatial server
```
`full` includes CellRank, FastCCC, the maintained spatial-domain and
registration backends, annotation, enrichment, and the other portable Python
families. R bridges, AESTETIK, and rctd-py remain separate because they have
system, platform, or large-runtime requirements. See the installation guide
before enabling those families.
Then:
1. **Run your first analysis** — [Quick Start](docs/quickstart.md)
2. **Choose optional method families or a persistent environment** — [Installation Guide](docs/installation.md)
3. **Configure another MCP client** — [Configuration Guide](docs/advanced/configuration.md)
4. **Reproduce the manuscript key results without an LLM** — [Reproduction image](reproducibility/README.md)
**Docker quick start:**
```bash
docker pull cafferyyang777/chatspatial:latest
```
**Minimal example prompt:**
```text
Load /absolute/path/to/spatial_data.h5ad and plot gene expression at the spatial coordinates
```
The [Quick Start](docs/quickstart.md) provides a downloadable sample and a
Leiden workflow that runs with the standard installation. Its sample has spatial
coordinates but no H&E image. SpaGCN requires the optional `spatial-domains` extra.
If you use Docker, mount host data to `/data` and prompt with the container path, for example `/data/spatial_data.h5ad`.
> ChatSpatial works with **any MCP-compatible client** — Claude Code, Claude Desktop, Codex, OpenCode, and other MCP-capable tools.
---
## Capabilities
Current coverage includes 67 methods across 13 analytical categories, exposed through 20 MCP tools. Supports 10x Visium, Xenium, Slide-seq v2, MERFISH, seqFISH.
| Category | Example methods |
|----------|---------|
| **Data Loading & Preprocessing** | Scanpy I/O, QC, Normalization, HVG, PCA, Neighbors |
| **Visualization** | Spatial plots, Embedding plots, Gene expression overlays |
| **Spatial Domain Identification** | SpaGCN, STAGATE, GraphST, BANKSY, AESTETIK, Leiden, Louvain |
| **Deconvolution** | FlashDeconv, Cell2location, RCTD (spacexr or rctd-py), DestVI, Stereoscope, SPOTlight, Tangram, CARD |
| **Cell-Cell Communication** | LIANA+, CellPhoneDB, CellChat (`cellchat_r`), FastCCC |
| **Cell Type Annotation** | Tangram, scANVI, CellAssign, mLLMCelltype, scType, SingleR, marker-score and threshold rules |
| **Differential Expression** | Wilcoxon, t-test, Logistic Regression, pyDESeq2 |
| **Trajectory Inference** | CellRank, Palantir, DPT |
| **RNA Velocity** | scVelo, VeloVI |
| **Spatial Statistics** | Moran's I, Local Moran, Geary's C, Getis-Ord Gi*, Ripley's K, Co-occurrence, Neighborhood Enrichment, Centrality Scores, Local Join Count, Network Properties |
| **Enrichment Analysis** | GSEA, ORA, Enrichr, ssGSEA, Spatial EnrichMap |
| **Spatially Variable Genes** | SpatialDE, SPARK-X, FlashS |
| **Multi-sample Integration** | Harmony, BBKNN, Scanorama, scVI |
| **CNV Analysis** | InferCNVPy, Numbat |
| **Spatial Registration** | PASTE, STalign |
---
## Documentation
| Guide | Use this when... |
|-------|------------------|
| [Installation](docs/installation.md) | You need optional methods or a persistent Python environment |
| [Docker](docs/docker.md) | You want a reproducible container runtime or local dependency resolution fails |
| [Configuration](docs/advanced/configuration.md) | You need exact MCP client syntax or the runtime path model |
| [Quick Start](docs/quickstart.md) | ChatSpatial is installed and you want the first successful analysis |
| [Concepts](docs/concepts.md) | You need to choose an analysis strategy from a biological question |
| [Examples](docs/examples.md) | You want copy-pasteable natural-language workflow prompts |
| [Methods Reference](docs/advanced/methods-reference.md) | You need canonical tool names, method names, parameters, and defaults |
| [Troubleshooting](docs/advanced/troubleshooting.md) | Setup, data loading, or analysis behavior is not working |
| [Full Docs](https://docs.cafferyang.com/) | You want the complete documentation site |
---
## Reproducibility
Each release publishes a reproduction image under the `-repro` tag. It
re-executes the DLPFC spatial-domain and OSCC CARD deconvolution tool calls
recorded during the manuscript experiments, without calling any LLM, and
compares the results with the reported values:
```bash
docker run --rm -v "$PWD/out:/outputs" \
cafferyyang777/chatspatial@sha256:ac3ba1ed8235c3d53ca72005d8ad2556cac007f5ddb932ae3e1d1241fa2cf9d8 \
chatspatial-reproduce
```
[`reproducibility/`](reproducibility/README.md) contains the image definition
and describes what it checks. The experiment and analysis scripts for the
manuscript are in
[ChatSpatial-Reproducibility](https://github.com/cafferychen777/ChatSpatial-Reproducibility).
---
## Citation
If you use ChatSpatial in your research, please cite:
```bibtex
@article{Yang2026.02.26.708361,
author = {Yang, Chen and Zhang, Xianyang and Chen, Jun},
title = {ChatSpatial: Schema-Enforced Agentic Orchestration for Reproducible and Cross-Platform Spatial Transcriptomics},
elocation-id = {2026.02.26.708361},
year = {2026},
doi = {10.64898/2026.02.26.708361},
publisher = {Cold Spring Harbor Laboratory},
URL = {https://www.biorxiv.org/content/early/2026/03/01/2026.02.26.708361},
journal = {bioRxiv}
}
```
ChatSpatial orchestrates many excellent third-party methods. **Please also cite the original tools your analysis used.**
---
## Contributing
Documentation improvements, bug reports, and new analysis methods are all welcome. See [CONTRIBUTING.md](CONTRIBUTING.md).
<div align="center">
**MIT License** · [GitHub](https://github.com/cafferychen777/ChatSpatial) · [Issues](https://github.com/cafferychen777/ChatSpatial/issues)
</div>
<!-- mcp-name: io.github.cafferychen777/chatspatial -->
TDQS
B3.4/5.0
Scored across 20 tools
Disambiguation4/5
Most tools target clearly distinct stages of spatial transcriptomics analysis, such as loading, preprocessing, embeddings, domain detection, deconvolution, and communication. A few pairs have overlapping scope—find_markers versus compare_conditions for differential expression, and analyze_spatial_statistics versus find_spatial_genes for spatial gene statistics—but descriptions and parameter differences help distinguish them.
Naming Consistency5/5
All tool names use snake_case and follow a verb_noun or verb_object convention (load_data, preprocess_data, find_markers, analyze_cnv, export_data). Verb choices vary appropriately by operation, with no inconsistent casing or mixed naming styles.
Tool Count4/5
The suite has 20 tools, slightly above the ideal 3-15 range, but this matches a comprehensive spatial transcriptomics workflow with many distinct analysis stages. Each tool earns its place, though the surface is somewhat heavy for an MCP server.
Completeness4/5
The tools cover the full analysis lifecycle: loading, preprocessing, embeddings, visualization, spatial statistics/domains/genes, deconvolution, annotation, cell communication, CNV, velocity, trajectory, and export/reload. Minor lifecycle gaps such as listing or deleting datasets prevent a perfect completeness score.
Maintenance
ActivityActive
ResponsivenessResponsive