The ultimate harness for coding agents and beyond. All your agents. All your machines. One command center. Start with code, then follow your curiosity and build across disciplines: CAD, circuits, robots, games and music.
TypeScript 3d-modelingagent-frameworkagent-harnessai-agentscadclaude-codecodexcoding-agentsdeveloper-toolsesp32game-developmentmcp-servermulti-agentopen-hardwareopencodeopenharnesspcbroboticsterminaltmux
import marimo
__generated_with = "0.24.2"
app = marimo.App(width="medium")
@app.cell
def _():
import marimo as mo
import numpy as np
import pandas as pd
import altair as alt
return alt, mo, np, pd
@app.cell
def _(mo):
mo.md(
"""
# A reactive notebook
Change the slider; every cell that depends on it re-runs. Tell the agent what you want to
analyse, model or show, and it rewrites this file — the pane follows.
"""
)
return
@app.cell
def _(mo):
n = mo.ui.slider(10, 500, value=120, label="points")
n
return (n,)
@app.cell
def _(n, np, pd):
rng = np.random.default_rng(7)
df = pd.DataFrame({"x": rng.normal(size=n.value), "y": rng.normal(size=n.value)})
df.describe()
return (df,)
@app.cell
def _(alt, df, mo):
chart = (
alt.Chart(df)
.mark_circle(size=60, opacity=0.7)
.encode(x="x", y="y", tooltip=["x", "y"])
.properties(height=320)
)
mo.ui.altair_chart(chart)
return
if __name__ == "__main__":
app.run()