What we are building
No dates, no promises. This is an honest view of what is live, what is actively in progress, and where the platform is heading.
Ask plain-language questions across disasters, demographics, economics, and risk. Get answers directly on a map without building a GIS workflow.
10 QA-gated packs covering 7 disaster categories, UN development indicators, global economic data, and more. Each pack passes source validation and coverage checks before release.
Browse packsMove across hazard, demographic, and economic layers in one conversation. Ask how disasters relate to population density, economic conditions, or development indicators.
Open the app without setup. No account required to start exploring. The starter catalog is available immediately.
Open the appAccount-based corpus management is live. Build a named corpus from published packs on your account page, load it in the app, and ask analytical questions against that data.
Learn moreThe engine is open source. Run your own instance, bring your own data, and use your own LLM key via GitHub.
View self-host docsAdditional source families are working through the QA pipeline. Demographic coverage, country-specific data, and expanded global indicators are in review.
See what is publishedDeepen the Research workspace with stronger memory, bounded artifact tools, and the ability to reopen and continue a prior analytical session.
User accounts that track access, credit balance, and which packs are available to a given session. The control layer for the hosted access model.
A focused live-conditions view for active events and operational monitoring. The map narrows to a geography or event and surfaces real-time signals alongside historical context.
Flight routes, shipping lanes, and satellite tracks as queryable data families. Designed to fit the same schema model as existing disaster and indicator packs.
GPU-accelerated overlays, particle effects, and higher-fidelity visual modes layered on top of the existing map without changing the data model.
Constrained versions of the engine for specific geographies, organizations, or problem domains. Same engine, reduced scope, custom catalog.
How this works
Packs go through a QA pipeline before they are published. Modes are designed before they are built, then promoted onto the public roadmap once the workflow is real enough to describe honestly. The roadmap reflects real internal priorities, not aspirational marketing.
Sign up to get notified when new packs, Research milestones, and local install updates ship.