mWater is the free, open-access data platform that 400,000 people use to map infrastructure, collect field data, and manage water and sanitation services in 198 countries. Since late 2025 it has had agentic AI built in, running in production. This page shows what we built and why it matters.
Every feature is free for every user, and no mWater feature has ever moved behind a paywall.
We spent fourteen years building the foundation AI needs to become useful: structured operational data, mobile field tools, organizational permissions, workflows, integrations, and a global community of public-sector users.
AI is now embedded inside that functioning system rather than sitting beside it as a chatbot. It removes specific technical bottlenecks, such as software integration, data analysis, and data cleaning, while users keep control of their systems.
mWater AI is deployed through the existing global platform, not as a separate product, so new capabilities reach users in 198 countries the day they ship.
Each one removes work that used to take hours or specialist skills. Together they have run more than 10,000 agent sessions, by users in 75 of the platform's 198 countries and territories.
Translate a survey into new languages with AI, then review every string before publishing. Teams deploy the same survey in the languages their enumerators and communities actually use.
How survey translation works →Ask for the chart, map, or table you need in plain language. An agent builds the widget, and the user can adjust everything by hand afterwards.
About mWater dashboards →An agent that connects mWater to external systems through our API. Users design the integration in a guided conversation instead of writing custom code, bringing sensors, gauges, and legacy databases into live operations.
The Integrator, explained →In June 2026 we released the first MCP server in the aid sector. Users connect the AI assistant they already use to mWater through the Model Context Protocol. Assistants can read our documentation, query data, and propose updates. A person approves every proposed change before it is applied.
Connecting AI assistants →“When we released the first MCP server in aid in June, LMICs jumped fastest to use it for good.”Dr. Annie Feighery, co-founder and CEO
Our AI proposes and people decide. The AI works inside your permissions, on your data, with your sign-off. That is what lets a public agency trust it with public records.
We publish how we think about AI. The short version: we treat AI agents as first-class users of the platform, and we build features that get better whenever the underlying models get better.
“AI doesn't make us design for superusers but stands to make everyone a superuser.”Petri Autio, Head of Product
Read the full AI strategy, and the product lead's essay on mWater as the backbone of the WASH sector.
These agents are not a pilot on the side. They run inside mWater, one of the most widely used data platforms in the water and sanitation sector, built on open source with an open API and no vendor lock-in.
operating continuously since, with new features released to all users at once.
platform uptime in 2024.
countries whose national monitoring systems are built on the platform.
of the platform's codebase is open source, with core libraries shared on GitHub.
languages the platform runs in.
countries and territories where people have already used the AI tools.
Explore the mWater Resource Center for documentation on everything the platform does.
A small team supports a platform of 400,000 users. AI helps our support staff answer questions from thousands of users while we keep an average response time of four hours.
Our engineers use AI coding agents daily. A small team ships a large platform, and every improvement reaches all users worldwide at once.
The mWater platform is a Digital Public Good. When a funder pays for a feature, we build it and release it to every user at no cost. An AI investment here serves the whole water and sanitation sector, not one project.
Our model costs to date are measured in hundreds of dollars, not millions. The scarce input is engineering time, and that is what funding buys.
mWater Foundation Inc. is a US 501(c)(3) nonprofit. Stellar Data, our commercial subsidiary, sells advanced support and data services and returns all profits to the Foundation.
If you fund AI for development, we would like to walk you through these features on real data and discuss what we could build next.
Building with AI? Connect your assistant through the mWater MCP server, or talk to Stellar about high-volume API access and technical assistance.
