Repository intelligence

hugohe3/ppt-master

An open-source local workflow that runs inside agent-capable AI tools to generate natively editable PowerPoint decks (PPTX) with real slide layouts, native shapes, and data-backed charts and tables from source documents.

CLOUDM0N decision

REVIEW BEFORE ADOPTION

Trust REVIEW · 70/100

Good fit if

Professionals who regularly build and refine presentations and want to automate the structural layout and drafting of PowerPoint slides without losing editability

Watch out for

Does not deliver a finished, perfect deck in one shot, requiring manual polishing and editing by the user

Practical intelligence

What matters before you adopt it

Decision first · detail on demand

Problem it solves

The tedious manual effort of drafting slide layouts and narratives, and the limitation of typical AI presentation generators that produce non-editable static images or simple filled-in templates rather than deep, natively editable PowerPoint files.

Best for

Professionals who regularly build and refine presentations and want to automate the structural layout and drafting of PowerPoint slides without losing editability

Developers and users of agent-capable AI tools (like Claude Code, Cursor, or Codex CLI) looking for an on-machine workflow to convert PDFs, documents, or raw text into decks

Main trade-offs

Does not deliver a finished, perfect deck in one shot, requiring manual polishing and editing by the user

Output quality is heavily dependent on the chosen AI model, with cheaper models causing a noticeable quality gap

SmartArt is deliberately omitted from the native PowerPoint feature mapping

Why it stands out

Maintains fully local data processing on the user's machine to protect sensitive source documents

Does not lock users into specific AI models or platforms, supporting Claude, GPT, Gemini, Kimi, and other agent-capable systems

Enforces real PowerPoint behaviors including native transitions, custom speaker notes, and opt-in animations

Trust & CVEs

Security evidence without the noise

Trust remains a decision signal; CVEs and scanner evidence explain what is driving the risk.

Security findings

CLOUDM0N scanner findings

Critical

High

Medium

Low

View trust evidence & security findings▼

Why this score

No trust rationale was stored for this scan.

CLOUDM0N findings

MEDIUM

Dynamic code execution pattern detected. 2 sample match(es) found.

MEDIUM

The project can spawn operating-system processes; review command construction and input handling. 5 sample match(es) found.

MEDIUM

Broad permission or elevated-command pattern detected. 1 sample match(es) found.

Architecture from code4 modules · 0 edges▼

Structural evidence

Modules and dependency edges extracted from repository code. This is code evidence, not README inference.

Code files

246

Modules

Dependency edges

Core modules

skills

246 files

Dependency flow

No dependency edges were extracted.

Detected languages

Python · JavaScript

Detected frameworks

Flask · NumPy · Requests

Architecture evidence details

flowchart TD
    %% repo — high-level architecture (DRAFT, refine me)
    n0["skills · 246 files"]

Evidence, security & integrations▼

Integrations

PythonPowerPoint (including DrawingML and OMML formatting)Kimi (Kimi K3 / Kimi Code)Claude (Claude Code, Claude Code CLI, Claude Code plugin)OpenAI (GPT, gpt-image-2, Codex CLI)Google Gemini (Gemini CLI, gemini-3.1-flash-image)Microsoft ResearchStudiocc-switch

Security notes

• Data stays local on your machine, ensuring source documents are not sent to third-party presentation platforms (only necessary API requests are sent to the configured AI model provider)

Still unknown

• The README does not provide concrete system hardware requirements (such as minimum RAM or CPU specifications) for running the local SVG-to-PPTX conversion

• Does not specify exact pricing or token consumption rates for generating an average slide deck on recommended models

• Does not detail how to configure slide masters manually or customize the styling guidelines for multi-language presentations

Adoption guidance▼

Adopt if

+ You want an open-source, local-first tool to generate fully editable .pptx files with native slide master layouts, shapes, and formulas

+ You are already utilizing agent-capable AI IDEs (such as Cursor or Claude Code) and want to integrate custom presentation workflows directly into your chat environment

+ You have strict data privacy requirements that prohibit uploading your materials to cloud-hosted AI presentation web platforms

Avoid if

– You expect AI to generate a polished, perfect, and ready-to-present deck in a single shot without any human review or touch-ups

– You are unwilling to set up Python, manage environment files, or learn to interact with agentic AI chat systems

– You are using unsupported presentation software (like Keynote, WPS, or LibreOffice) and require complex math formulas to render and edit flawlessly on those platforms

How it works & getting started▼

How it works

1.The user installs Python 3.10+, downloads the ppt-master project, and opens the folder in an agent-capable AI tool

2.The user places their source documents (PDF, DOCX, text) in the projects/ directory and prompts the AI agent to make a deck

3.The AI analyzes the content, reasons the argument into shape, and drafts a design specification

4.The AI designs the slides, generates corresponding SVGs, and runs Python post-processing scripts

5.An integrated PPTX converter reads the generated SVGs and exports a natively editable, structured .pptx file in the exports/ directory

Getting started

→ Download Python 3.10+ (ensuring 'Add to PATH' is checked during installation)

→ Clone the repository or download the ZIP from GitHub, and run 'pip install -r requirements.txt' inside the project folder

→ Copy the template environment file using 'cp .env.example .env' and add your API keys

→ Open the 'ppt-master' directory in your chosen agent-capable AI tool (like VS Code or Claude Code) and launch the chat

Agent handoff

Use with any agent

JSON API

Alternatives

Nearby repositories worth comparing before adoption.

Compare top options →

slidev

slidevjs/slidev

Fit

Presentation Slides for Developers

Trust REVIEW · 70

Compare →

MinerU

opendatalab/MinerU

Fit

Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.

Trust REVIEW · 80

Compare →

docling

docling-project/docling

Fit

Get your documents ready for gen AI

Trust FAIL · 0

Compare →