Python Office Automation That Lands Live, Attributed

The Athena Python SDK is a set of drop-in packages from Athena Intelligence for Python office automation. Install athena-python-pptx, athena-openpyxl or athena-python-docx, keep the same imports, and open a live deck, workbook or document by its asset ID. Edits land in the shared file your team is working in, attributed to the API key's owner.

Most data and engineering teams already have the scripts: a python-pptx job for the Monday pipeline review, an openpyxl job for the regional numbers, a python-docx job for the memo. Each one ends with a file on disk that somebody still has to get in front of the team.

Athena Intelligence, founded in 2022, builds an agent platform for regulated enterprises where people and agents build agents, workflows and applications together. The Python SDK lets existing office code write straight into the live files in Athena's studios. Figure 1 shows the whole idea on one page.

• The imports stay the same. The code opens a live asset by its ID.
• Deck edits are sent call by call as the script runs, or grouped with prs.batch(). Workbook edits are buffered until wb.save().
• Every edit carries the API key owner's email or agent ID, on the shape and in the Activity pane.
• The launch video runs 30 seconds. Every name and number in it is demo data.

01Why does Python office automation stop at a file?

Python office automation stops at a file because python-pptx, openpyxl and python-docx are file libraries. Each one writes to disk, and the shared copy the team works in lives somewhere else.

A typical pipeline-review script ends with prs.save("q3_review.pptx"). Then someone uploads, emails or pastes the result into the deck the team is already editing. If a colleague changed slide 1 that morning, two versions now exist, and one author field covers the whole file. Figure 2 sets the two paths side by side.

That shared copy is usually in the cloud. In 2025, 52.7% of EU enterprises bought cloud computing services, and 71.7% of those used cloud office software. Governance for automated work lags behind: in a 2025 survey of 360 IT application leaders, only 13% strongly agreed they had the right governance structures to manage AI agents. A script writing into a shared file raises the same question: who changed what, and when.

02How does the Athena Python SDK work?

You install Athena's package for the library you already use, keep your imports, and pass an asset ID where the code used to pass a file path. Athena rewrote the three office libraries on a CRDT layer and published them on PyPI.

A CRDT (conflict-free replicated data type) is a data structure that lets several writers change the same document and merges their changes, so a script and a person can work in one deck at once.

pip install athena-python-pptx athena-openpyxl athena-python-docx

The deck script from the launch video, pointed at a live deck, looks like this. Everything after the Presentation line is ordinary python-pptx.

from pptx import Presentation
from pptx.util import Inches
from pptx.chart.data import CategoryChartData
from pptx.enum.chart import XL_CHART_TYPE

prs = Presentation(asset_id="asset_7f3c2a91")  # a live Athena deck
slide = prs.slides[0]

slide.shapes.title.text = "Q3 pipeline review"
box = slide.shapes.add_textbox(
    Inches(0.6), Inches(1.5), Inches(6), Inches(0.5))
box.text_frame.text = "Pipeline up 18% on Q2, led by EMEA"

data = CategoryChartData()
data.categories = ["NA", "EMEA", "APAC", "LATAM"]
data.add_series("Q3", (4.2, 3.1, 1.8, 0.9))
slide.shapes.add_chart(
    XL_CHART_TYPE.COLUMN_CLUSTERED,
    Inches(0.6), Inches(2.5), Inches(12), Inches(4.6), data)

Each call is sent to the live deck as the script runs, so anyone with the deck open sees the title, then the text box, then the chart. To group calls, wrap them in prs.batch():

with prs.batch():
    slide.shapes.title.text = "Q3 pipeline review"
    box.text_frame.text = "Pipeline up 18% on Q2, led by EMEA"

The script runs with an Athena API key, and the key's owner is the name on every edit. Figure 3 compares the three packages.

03What does the launch video show? One example, start to finish

The 30-second video follows one python-pptx script from install to a finished slide, then repeats the pattern for a workbook and a document. Figure 4 lists the six steps with their captions.

1. Install. pip install athena-python-pptx ends with "Successfully installed athena-python-pptx-0.13.1". The script, q3_review.py, builds a Q3 pipeline review slide.
2. Point. The imports stay as they were. One line changes, from Presentation("q3_review.pptx") to Presentation(asset_id="asset_7f3c2a91").
3. Run. The title, the text box "Pipeline up 18% on Q2, led by EMEA" and a column chart for four regions appear in the deck one after another.
4. Watch. A teammate, Morgan Lee, is in the deck, and the run's avatar reads "dana@acme.example · dana@acme.example's run".
5. Attribute. The Timeline tab in the Activity pane lists each change under Dana Ruiz, dana@acme.example.
6. Extend. The same pattern writes a row to a workbook and a heading and paragraph to a memo.

The end card reads "Keep your python-pptx code. Every edit lands live, attributed."

Table 1 compares the file-based script with the same script on the SDK.

04Does it work for workbooks and documents?

Yes. athena-openpyxl and athena-python-docx follow the same pattern: install the package, keep from openpyxl import Workbook or from docx import Document, and open the live asset by ID.

from openpyxl import Workbook

with Workbook("asset_9b21e4c0") as wb:
    ws = wb.active
    ws.append(["EMEA", 3.1, 0.18])
    wb.save()

Workbook edits are buffered and land in the shared workbook when the script calls wb.save(). Inside the block, the calls are the ones openpyxl users know, such as wb.active and ws.append(). Athena's sheets have handled 95 sub-agents writing to one sheet without conflicts.

from docx import Document

with Document("asset_c4d81f2a") as doc:
    doc.add_heading("Q3 pipeline review", level=1)
    doc.add_paragraph("EMEA led the quarter.")
    doc.save()

The document script adds a heading and a paragraph with the usual python-docx calls, then saves.

05How is each edit attributed?

Every edit made through the SDK is attributed to the owner of the API key that ran the script. The name shown is the key owner's email or an agent ID.

While the script runs, people in the deck see a coloured ring and a name label on each shape as it changes, a pulsing avatar with the tooltip "name · email's run", and a presence dot on the slide. Afterwards, the Activity pane keeps the record in its Attribution, Timeline, Sessions, Audit and Time Travel tabs, and each change reads "Edited by" and the name. Figure 5 draws both views.

Across the platform, Athena can trace cell-level attribution to the exact tool call within seconds, and attribution and rollback cover about 100 data types. Since the name comes from the key, decide who owns the key a scheduled job uses before the job goes live.

06Is Athena secure enough for regulated companies?

Yes. Athena Intelligence's security program includes SOC 2 Type II and HIPAA, third-party penetration tests and zero data retention agreements with frontier model providers, and Athena does not train on customer data.

Athena deploys in a managed cloud, in your own VPC on AWS, Azure or GCP, on-premises, or air-gapped, including US government air-gapped networks and AWS GovCloud. A Big Four audit firm has run Athena in production for over a year, with over 1,000 users.

07How do we get started?

Pick one script your team runs every week, such as the Monday deck builder. Install the matching package, swap the file path for the asset ID of a deck the team already uses, and run the script while someone has the deck open.

The SDK and API page lists Athena's other developer surfaces: the athenaintel Python client, a TypeScript SDK, React components and a public API with run budgets. Dispatch No. 12 covers people and agents editing the same file.

Curious how it would work on your scripts? We're happy to show you.

08Questions teams ask