PyGWalker HTML Export API

Use pygwalker.to_html(...) when you need a static HTML string. Static exports are browser-only: they do not support live kernel or cloud computation.

import pygwalker as pyg
 
html = pyg.to_html(
    df,
    spec_path="./gw_config.json",
    computation="browser",
)

Static HTML limitation

Static HTML has no live Python kernel and no cloud callback. PyGWalker rejects:

pyg.to_html(df, computation="kernel")
pyg.to_html(df, computation="cloud")
pyg.to_html(df, kernel_computation=True)
pyg.to_html(df, cloud_computation=True)
pyg.to_html(df, use_kernel_calc=True)

Use computation="browser" or omit the computation option. Use Notebook APIs, Streamlit, Gradio, or webserver mode when you need live kernel or cloud computation.

`pygwalker.to_html`

to_html returns an embeddable iframe HTML string.

import pathlib
import pygwalker as pyg
 
html = pyg.to_html(df, spec_path="./gw_config.json", computation="browser")
pathlib.Path("pygwalker.html").write_text(html, encoding="utf-8")

Signature:

pyg.to_html(
    df,
    gid=None,
    *,
    spec="",
    spec_path=None,
    field_specs=None,
    theme_key="g2",
    appearance="media",
    default_tab="vis",
    computation=None,
    **kwargs,
) -> str

df can be a pandas DataFrame, polars DataFrame, pyarrow Table, or reusable pygwalker.Walker.

Export from a reusable `Walker`

Create a Walker when the same chart state should also render in notebooks or apps.

import pygwalker as pyg
 
walker = pyg.Walker(
    df,
    spec_path="./gw_config.json",
    computation="browser",
)
 
html = pyg.to_html(walker, width="100%", height="720px")

When to_html receives a Walker, pass rendering dimensions as width and height keyword arguments. Construction options such as spec_path, field_specs, appearance, and computation must be set on the original Walker.

walker.to_html(iframe_width=None, iframe_height=None) and walker.to_html_without_iframe() provide the same static-export behavior and also reject live kernel/cloud computation.

Table and renderer helpers

Use these helpers when you do not need the full explorer.

from pygwalker.api.html import to_table_html, to_render_html
 
table_html = to_table_html(df, spec_path="./gw_config.json", computation="browser")
render_html = to_render_html(df, spec_path="./gw_config.json", computation="browser")

Signatures:

to_table_html(
    df,
    *,
    spec_path=None,
    theme_key="g2",
    appearance="media",
    computation=None,
    **kwargs,
) -> str
 
to_render_html(
    df,
    spec="",
    *,
    spec_path=None,
    theme_key="g2",
    appearance="media",
    computation=None,
    **kwargs,
) -> str

Chart preview helper

to_chart_html renders a single chart from a Graphic Walker or Vega spec.

from pygwalker.api.html import to_chart_html
 
chart_html = to_chart_html(
    df,
    spec=chart_spec,
    spec_type="graphic-walker",
)

Signature:

to_chart_html(
    dataset,
    spec,
    *,
    spec_type="graphic-walker",
    theme_key="g2",
    appearance="media",
) -> str

dataset can be a DataFrame, pyarrow Table, database Connector, or connector-style string.

Common traps

Trap	Fix
Exporting large local data with kernel computation	Use a live backend instead of static HTML.
Passing a `Walker` plus `spec_path` to `to_html`	Put `spec_path` on `pyg.Walker(...)`.
Starting new static examples with `kernel_computation=True`	Use `computation="browser"`.
Expecting users to edit and save chart state in a static export	Use a live notebook/app backend with `spec_io_mode="rw"`.

Related Guides

Gradio Component API