Work in SQL and YAML? Switch to the engineer view →
Business intelligence as code

Every number defined once. Every change reviewed.

Define each metric once and every dashboard uses it. Every change waits for a commit before it goes live, whether it comes from the visual UI, an AI agent or code.

✓ $30 of credit✓ No card✓ Unlimited members
“The ease of developing and maintaining dashboards has resulted in our team quickly building a much higher degree of visibility into our data.”
Dan Manges, Co-Founder & CEO at RWX
Three ways to work, one source of truth
ConnectSource
ShapeModel
DefineMetric
ExploreInsight
VisualizeChart
ShareDashboard
The problem with BI today

When nobody trusts the numbers, nobody acts on them.

Visivo is open-source BI as code. You define each number once, and people and coding agents build dashboards on it. Every change is a draft you review before it ships.

✕“Whose number is right?”

One definition of every number

Declare metrics, dimensions and joins once. Every chart reads the same definition. Rename a metric and see every chart it touches first. Ten databases, with or without dbt™.

Semantic layer + lineage · live
✕“Can you trust the AI?”

Agents draft, people decide

Claude Code, Cursor or the built-in agent edits your project over MCP. Every write lands as a draft. Nothing reaches your files or dashboards until a person commits it.

MCP + drafts · live
✕“Dashboards keep breaking?”

It ships like software

Browser, files or agent, it all saves to the same YAML. Publish with visivo deploy, or export a static site that still filters.

BI-as-code + visivo dist · live
How Visivo works

From raw source to trusted dashboard.

Every object in Visivo has its place in one connected flow, the same colors you'll see inside the product's lineage and explorer.

  1. 1ConnectSourceConnect 10 data sources + dbt™. Credentials stay in your infrastructure or a write-only secret store.
  2. 2ShapeModelTransform raw data into clean, version-controlled SQL models.
  3. 3DefineMetricDefine each number once in the semantic layer.
  4. 4ExploreInsightAggregate once, then filter, split & sort in the browser.
  5. 5VisualizeChartRender 48 chart types from the same trusted data.
  6. 6ShipDashboardCompose, deploy & govern with review and lineage.
Three ways to work

One source of truth. Worked three ways.

Build in the visual UI, ask an agent, or write the YAML by hand. Agents draft; you review and commit. All three work on the same version-controlled project and semantic layer. Pick one per task, not per person.

Visual UI: drag chart tiles onto a canvas; Visivo writes the layout into project.visivo.yml.

You drag tiles onto a live canvas, and Visivo writes the layout into the same project.visivo.yml.

project.visivo.ymlone source of truth
metric net_revenue: …chart mrr_by_month: …dashboard Revenue: …
One semantic layer · one Git history · type-checked against a JSON schema
See it work

From a SQL query to a dashboard your team trusts.

Dashboards

Run visivo deploy. Your whole team opens the dashboard.

KPI tiles and charts on one page, built in a project your team reviews like code.

  • 48 chart types, from candlesticks to 3D scatter
  • Teammates sign in from a browser, with roles from Viewer to Admin
  • Only the builder needs the CLI
EV Sales Overview dashboard in Visivo Cloud with two KPI tiles and three charts
Units sold, revenue, units by quarter and region, and the BEV and PHEV split on one page your team opens in a browser (sample data).
Explore

Write SQL, then drag columns onto a chart.

Run a query against your database and drag result columns onto x and y. When the chart looks right, save it to the project.

  • Profile any column: distribution, nulls and distinct values
  • Pick a chart type and it lists the fields it needs
  • Reuse the saved chart on any dashboard
Running a SQL query in Visivo and dragging two result columns onto a bar chart
Run the query, drag quarter to x and units to y, then title the bar chart, all on one screen (sample data).
Lineage

Flip any dashboard to see what its charts depend on.

See the source, model and charts behind it, plus the metrics defined on that model. Know what a change will touch before you make it.

  • Models and metrics sit in the same graph as the charts
  • Switch between Canvas and Lineage on any dashboard
  • Filter the graph by source, model or chart name
A dashboard flipping to a graph of its source, model, metrics and charts
Flip the EV Sales dashboard to its source, model and charts, then show the full project to add the metrics (sample data).
Review and commit

Edit in the browser. Nothing goes live until you commit.

Change a dashboard in Visivo Cloud and run the project to check the result. Then commit it with a message, the same way you ship code.

  • Your edits stay in a draft until you commit
  • Each run shows whether it worked, with the full log
  • Link a GitHub repo and a commit can open a pull request
Changing a dashboard row height in Visivo Cloud, running the project, then committing with a message
Change a row height, run the project to check it, then commit with a message (sample data).
Trusted by data teams

Teams ship faster when they trust the numbers.

“The ease of developing and maintaining dashboards has resulted in our team quickly building a much higher degree of visibility into our data. It's a huge improvement in security over giving a tool unfettered access to databases.”
Dan MangesDan Manges
Co-Founder & CEO at RWX
“Visivo made it easy for us to deliver fast, reliable dashboards that our stakeholders can trust. Having a single source of truth, with full version control, has completely transformed the way our team builds and shares insights.”
Clara DavisonClara Davison
Analytics at Empora
Built by people who've done itThe team scaled analytics at:
IntuitBoeingRoot Insurance
Open source · BI-as-codeInspect it, fork it, build on it. Star on GitHub
The strongest proofLive, clickable example dashboards. Explore live examples
Install command copied