A shared space for data

Just ask Neo.
Run it locally.

Tell Neo what you need. Local LLMs turn natural language into data workflows, and our C++ engine runs them on your machine. Clean, join, and chart your data automatically, with the freedom to inspect and change every step. Keep your work local. Share what you choose.

Open your workspace Follow the flow
Cloud + desktop Work together, live
Revenue workspaceExample workflow

A shared view of the work. Coordinate the workflow and review its results in your browser.

Open cloud workspace
Different screens. The same shared work.Illustrated team roles
TylerBrings in the data
Desktop
IsabelConnects the tables
Desktop
CodyBuilds the workflow
Desktop
JoelReviews the data
Cloud
SilasShares the charts
Cloud
AntoineWorks with Neo
Desktop

An example of the work coming together. Everyone contributes; each person has their own view.

From source to shared result

One workflow,
from source to chart.

See the joins, filters, and calculations behind every chart. Build the next step with your team.

01 / Bring it in

Start with your data.

Bring files and sources into the workspace. Organize what you need and choose what to share with your team.

Explore your data space
02 / Make it useful

Clean. Join. Make it yours.

Connect tables, filter rows, and add calculations in ET1. Follow every step on the canvas, or ask Neo to help.

Explore workflows
03 / See the result

Give the answer a shape.

Turn the finished data into charts and dashboards. Arrange the view, refine the details, and share what matters.

Explore dashboards

Different screens. Shared work.

Cloud or desktop.
Still on the same team.

A place to bring the people in, and room for everyone to get to work.

In the cloud

Keep the team connected.

Invite people into a shared room. Share workflows and results, discuss the next step, and keep the work in one place.

Open Collaborate
On your desktop

Go deep with your data.

Neo helps you work across your data, ET1 workflows, Rug, and charts using models on your machine. Keep your offline tools, and join your team when you’re ready.

Explore the desktop canvas

Made to be shared

Embed a chart.
Just add code.

Take a published chart from your workspace to your website. Copy one tag and give the result a place to live.

The embedded resultPublic benchmark

Sum by region

Lower time is faster · milliseconds

Compared engineneo in the same run
  1. Zig0.024 ms

    neo in this run 0.020 ms

  2. Swift0.031 ms

    neo in this run 0.020 ms

  3. Rust0.042 ms

    neo in this run 0.020 ms

  4. Go0.079 ms

    neo in this run 0.020 ms

  5. Polars1.546 ms

    neo in this run 0.020 ms

  6. DuckDB2.239 ms

    neo in this run 0.020 ms

  7. pandas12.239 ms

    neo in this run 0.020 ms

  8. Spark38.959 ms

    neo in this run 0.020 ms

One scale across all rows. Tiny bars are kept visible.

This is the published sum by region benchmark, rendered by the actual embed viewer.

The same public benchmark data, arranged for your phone. Open the full chart beside the embed code.

Your chart.
One line of code.

JavaScript
<script async src="https://n0.is/embed.js" data-src="https://n0.is/embed/chart/sum" data-title="sum by region"></script>
Open chart

Ready for your page. This example uses a fixed public benchmark snapshot from n0.is.

Want it to update?

Connect local output to an explicitly published live result. The optional connector watches your exported readout and sends changed display data.

01

Choose what to share.

Publish the display data you choose. Your source workflow and connection credentials stay private.

02

Send changed results.

The local connector sends an update only when the chart data changes.

03

Keep control.

Publish explicitly. Revoke a live grant to remove its result from open viewers, too.

04

Use the work again.

Embed an already computed chart or dashboard. Viewing it does not rerun your source aggregation.

Under the surface

A clear workspace.
A capable engine.

Explore the published engine comparisons, exact-output checks, and the cases where another engine is faster.

See the measurements
Ahead in published comparisons
146 / 150

Head-to-head measurements

Over the JavaScript it replaces
33×

Median across measured kernels

Same output, old and new
627 / 627

Exact-output verification cases

There’s room for your next idea

Bring someone
into the flow.

Your data, your questions, and the people who help you make sense of them.

Open your workspace Let’s talk