Mprove
Open Source Agentic Business Intelligence with Malloy Semantic Layer·mprove.io
Mprove is an open-source business intelligence platform built around the Malloy semantic layer. It lets teams discover data warehouse schemas, enrich metadata, and build reusable semantic models, then explore data through an AI data analyst and build reports and dashboards. Changes to models and dashboards are deployed as code using Git version control, with support for personal developer repositories, a shared production repository, and multiple environments (Test, Stage, Prod).
What it's for
Features 18
- AI data analyst available for self-service data answers
- Concise Malloy queries that translate into SQL
- Drill-down segmentation of metrics by dimension
- DWH schema discovery (tables, columns, indexes, keys, data types)
- Git commands for managing changes (commit, push, pull, fetch, merge, revert)
- Git version control with personal dev repositories and a shared production repository
- Global filters linked across dashboard tiles or report metric rows
- HTTP API connectors including Custom API and Google Analytics
- Instant validation of model changes on save
- LLM provider and model configuration for AI Explorer and Editor sessions
- Metadata enrichment on tables and columns (description, example data, relationship type)
- Multiple environments for querying (Test, Stage, Prod)
- Reusable semantic models built with Malloy (DRY SQL blocks)
- Role-based access control for models, reports, and dashboards
- Sandbox provider support (E2B) for running AI agent sessions
- Simple UI for building reports and composing dashboards from multiple charts
- SQL database integrations (PostgreSQL, MySQL, Databricks, Snowflake, BigQuery, MotherDuck, Presto/Trino)
- Visual data relationships graph with click-to-sample data
At a glance
Integrations
Resources
Pricing
Mprove is self-hosted and open source; no subscription pricing is listed. A login-free live demo with limited features is available without self-hosting.