A conversation. A notebook. Your analysis.

Chat with your data.
Keep the analysis.

Ask in plain English. Hailer explores your files and builds tables, charts and editable Python in a live marimo notebook.

Get started See it in action

CSV · Parquet · JSON · Excel / Windows, macOS & Linux

From question to notebook

Sample walkthrough · sales.csv

hailer / my-analysisTERMINAL

$ uvx hailer

Notebook ready. What would you like to explore?

YOU

Chart total revenue by month,
split by region.

HAILER

I've added a chart to your notebook. Both regions grew from January to February.

Chart added to analysis.py

Ask a follow-up…

analysis.py · conversation saved

analysis.pyMARIMO NOTEBOOK

SALES OVERVIEW

Small file. Clearer picture.

sales.csv

NorthSouth

Revenue by month, split by region. Four rows, two regions.

THE CELLS BEHIND THE CHART

Readable. Editable. Yours.

import polars as pl
import altair as alt

sales = pl.read_csv(DATA_DIR / "sales.csv")
revenue = sales.group_by(
    "month", "region"
).agg(pl.col("revenue").sum())

alt.Chart(revenue).mark_bar().encode(
    x="month:N", y="revenue:Q",
    color="region:N", xOffset="region:N",
)

Illustrated with sample data. Download the CSV and try it yourself.

MORE THAN A CHAT HISTORY

Something to come back to.

Follow the next question. Your notebook keeps the work.

01 / EXPLORE

Start with your files.

Find missing values, compare months or explore an Excel worksheet. Ask the question in the words you'd use with a colleague.

Working with your data

02 / UNDERSTAND

See how you got there.

Charts and tables sit alongside their Python code. Open the cells, inspect the calculation and make your own changes in marimo.

The everyday workflow

03 / CONTINUE

Pick up the thread.

Reopen your workspace to resume the conversation and active notebook. Start another notebook when the analysis moves somewhere new.

Notebooks and sessions

YOUR FIRST ANALYSIS

Bring a question. Start here.

You'll need uv, a browser and a model API key.

1
Create a workspace
A folder for your data, notebooks and settings.
```
mkdir my-analysis
cd my-analysis
uvx hailer init
```
2
Connect your model
Sign in with an OpenAI API key, or configure another endpoint.
```
uvx hailer login openai
```
3
Add data. Say hello.
Put files in data/, then start Hailer. Keep the notebook tab open.
```
uvx hailer
```

Follow the complete quickstart

KNOW WHERE YOUR DATA GOES

Local files.
A model you configure.

Your analysis runs on your machine. Your messages, code and notebook tool outputs — which can include data samples — go to your configured model endpoint.

By default notebook code runs in a Docker container that sees only your data folder, read-only, with no network; its outputs still go to the model. Running code as you, with your files and network, is an explicit opt-in (unsafe-local).

Read about data handling Choose your runtime

TAKE IT FROM HERE

Make Hailer part
of your workflow.

Teach it your projectContext, reusable prompts and skills Find a commandTerminal commands and chat shortcuts Get unstuckDiagnostics and fixes for common problems Explore the projectArchitecture, development and tests

Chat with your data.Keep the analysis.

From question to notebook

Small file. Clearer picture.

Readable. Editable. Yours.

Something to come back to.

Start with your files.

See how you got there.

Pick up the thread.

Bring a question. Start here.

Create a workspace

Connect your model

Add data. Say hello.

Local files.A model you configure.

Make Hailer partof your workflow.

Chat with your data.
Keep the analysis.

Local files.
A model you configure.

Make Hailer part
of your workflow.