Reproducibility-first AI infrastructure for computational hydrology and Earth system science.

Fewer than 7% of published computational hydrology studies provide sufficient documentation for independent replication. AI-Hydro exists to change that — by making reproducibility a natural byproduct of the research process, not an afterthought.

We build open, AI-native infrastructure that lets hydrologists describe their intent and get real, documented, citable computation back — without writing glue code or losing provenance.

An AI agent for hydrological research built on the Model Context Protocol. Give it a USGS gauge ID and a research question — it delineates watersheds, retrieves streamflow, extracts signatures, calibrates models, and records every step with structured provenance.

pip install aihydro-tools[all] · Documentation · VS Code Marketplace

The Python backbone powering the AI-Hydro platform. A suite of hydrological and geospatial analysis tools exposed via MCP — usable from any AI model (Claude, GPT, Gemini) or directly as a Python library.

pip install aihydro-tools · PyPI · Tool Reference

An optional semantic search layer for hydrological literature using vector embeddings. Archived while the simpler folder-based literature module matures in the core platform.

Researcher (natural language)
        ↓
  AI Agent (Claude / GPT / Gemini)   ← any MCP-compatible model
        ↓  JSON-RPC over stdio
  aihydro-mcp (Python / FastMCP)
        ↓                    ↓                      ↓
  Federal APIs         ML backends          ~/.aihydro/
  USGS · GridMET       conceptual +         HydroSession
  3DEP · NLCD          deep learning        ProjectSession
  NHDPlus · CAMELS                          ResearcherProfile
                              ↓
                      HydroResult { data, meta }
                      — source, parameters, timestamp
                      — automatic provenance, every step

AI-Hydro is a community platform. The most impactful contributions are new domain tools — flood frequency analysis, sediment transport, groundwater modelling, remote sensing, snow hydrology — packaged as Python entry-point plugins.

You don't need to fork the core. Write a package, register an entry point, publish to PyPI. It's immediately available to every AI-Hydro user.

High-priority domains open for contribution:

→ Plugin Development Guide · Open an issue

_{Built with ❤️ for the hydrology and Earth system science community · Apache 2.0}