tessera

Build self-contained, interactive HTML reports from Python — one file you can commit, or serve, with no runtime and no dependencies on the viewer's machine. Designed for batch-generated data and ML output: loop over your experiments, runs, or segments and emit a report per iteration.

Content is laid out on a grid of typed cells (charts, tables, metrics, code, images, diagrams), and the whole deck embeds into a single .html.

Documentation

https://tessera-report.readthedocs.io/

Installation

pip install tessera-report

Quick start

from tessera import Deck

deck = Deck(
    title="My Report",
)

deck.add_title("My Report", subtitle="Subtitle here")
deck.add_section("1 — Introduction")

slide = deck.add_slide("Results", ncols=2)

slide.add_metric(value=98.7, label="Efficiency (%)", delta=+2.3)
slide.add_text("Text with **markdown** and LaTeX: $E = mc^2$")

deck.write("report", open_browser=True)

Batch generation

The core use case — one report per run, generated in a loop:

deck = Deck(title=f"Experiment Results", author="Company name", date="Jun 20th, 2026")

for run in experiment_runs:
    
    deck.add_section(title=f"Run #{run.id}")    
    
    slide = deck.add_slide("Metrics", nrows=1, ncols=3)
    
    slide.add_metric(value=run.accuracy, label="Accuracy", delta=run.delta)
    slide.add_plotly(run.loss_curve)
    slide.add_table(run.results_table, caption="Results")

deck.write("report")

Cell types

Common cell options

All add_* methods accept:

slide.add_text(
    "content",
    colspan=2,          # spans 2 columns
    rowspan=1,
    caption="Caption",
    overflow=True,      # internal scroll
    expand_button=True, # expand to fullscreen button
    copy_button=True,   # copy button (CodeCell)
    transparent=True,   # hides cell border and background
)

License

MIT