A data visualization platform built for AI adoption, giving LLMs the tools to turn project data into clear decision-ready dashboards.
A set of styled, ready-to-use project-controls dashboard templates a team can run on its own data — upload existing Excel logs, or link a OneDrive or SharePoint folder, and d3bi renders cost/EVM, schedule, RFI/submittal, risk, and quality boards.
The charts are authored from raw D3 math — d3-scale, d3-shape, d3-axis — through React JSX rather than composed from a chart library. Now shipped as a login-gated app in production, validated against two real Sound Transit East Link (2 Line) packages, E130 and E320, at a 31 March 2021 data date.
Every number traces to a validated earned-value engine — never to model prose.
debi — the dashboard design assistant
An in-app AI assistant turns a plain-language request into a complete, data-validated dashboard and saves it as a reusable template. It reads what a team's uploaded data actually supports through a readiness loop and emits panels, KPIs, and layout that render live, never fabricating a metric. Allows executives to draft reusable dashboards fast, without getting buried in measures, mappings, or data modeling.
MCP & CLI — bring your own inference
A custom MCP server and CLI let a customer drive d3bi from their own LLM subscription — connect Claude Code, Codex, or a claude.ai connector and their subscription pays for the inference, so d3bi costs only hosting and the database. Both run on the exact tool layer the in-app assistant already uses: additional front doors onto one backend, not a second system.