PPTAgent: An Agentic Pipeline for Reflective Slide Generation
An Agentic Framework for Reflective PowerPoint Generation
At a glance
- What is it?
- PPTAgent is a Python framework for reflective PowerPoint generation, now shipped as a Skill that lets Claude Code, Codex CLI and OpenCode create, revise and visually review editable decks. It suits researchers and engineers who need structured, content-dense slides but not pixel-perfect design control.
- Who is it for?
- Adopt PPTAgent if you already work in Claude Code, Codex CLI or OpenCode on Linux, WSL or macOS and want your coding agent to create, revise and visually review editable PowerPoint decks. Do not use it if you require fine-grained visual design or Windows-native support.
- Can I use it commercially?
- Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
- Is it still maintained?
- Yes. The repository last received commits 6 days ago.
- What is it written in?
- Mainly Python, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on November 13, 2026, and from our analysis. They are not legal advice.
Star history
- icip-cas/PPTAgent5,244 ★ · +229
| Repository | 2026-09-10 | 2026-11-13 |
|---|---|---|
| icip-cas/PPTAgent | 5k | 5.2k |
229 Stars in the last 64 days, measured from the daily GitHub snapshot. Compare with other projects
Editorial analysis
What PPTAgent Actually Solves
PPTAgent addresses the gap between generic text-to-slide tools and the manual labor of structuring a presentation. It is built for people who need to turn a research paper, a report, or a set of data files into a slide deck where each slide has a clear purpose and the overall narrative flows. The project's own abstract mentions a 'reflective' generation process, which suggests that the agent does not just dump content into templates but iterates on structure. The target user is an engineer or researcher who already works with LLMs and wants to automate the first draft of a presentation, not a designer who cares about pixel-level aesthetics. The README positions it as an 'Agentic Framework', meaning it orchestrates multiple steps and tools rather than running a single model call. This is a practical fit for academic and corporate settings where the bottleneck is content organization, not slide decoration.
The Reflective Three-Stage Mechanism
For the method itself, the README points to the two papers: PPTAgent (EMNLP 2025) at arxiv.org/abs/2501.03936 and DeepPresenter (ACL 2026) at arxiv.org/abs/2602.22839. What the current branch describes is a Skill that lets a coding agent create, revise and visually review editable PowerPoint decks, with an optional multimodal model reviewing the rendered slide images.
DeepPresenter: The Fine-Tuned Model That Changes the Deal
DeepPresenter is the second research line in this repository. Its codebase was released in 2025/12 with Deep Research Integration, Free-Form Visual Design, Autonomous Asset Creation, Text-to-Image Generation and an Agent Environment with sandbox & 20+ tools, and the v2.0.0 release is titled "PPTAgent V2: DeepPresenter Integrated". Fine-tuned models and a taskset were released on Hugging Face in 2026/03, and the DeepPresenter paper was accepted to ACL 2026. Its code is pinned at v1.1.38, published 2026-09-20, rather than on the current development branch, which now centres on the Skill.
Getting It Running: CLI, Docker, and Configuration
The current install path is the Skill. It requires Claude Code, Codex CLI or OpenCode; Linux (including WSL) or macOS; uv; npm; and LibreOffice available as `libreoffice` on PATH, plus Google Chrome on macOS for the converter. On Debian/Ubuntu you run `sudo apt-get install npm libreoffice-impress`; on macOS you install node, libreoffice and google-chrome with Homebrew. Then clone the repository, move into `PPTAgent/skills/pptagent`, run `uv venv --python 3.12 .venv`, `uv pip install --python .venv/bin/python -r requirements.txt` and `.venv/bin/python -m playwright install --with-deps chromium`. Registration is one command per client, `.venv/bin/python scripts/install.py --client claude` (or `--client codex`, `--client opencode`), and `.venv/bin/python scripts/pptagent.py doctor` checks the installation. Keep the repository in place afterwards.
Optional Services That Shape Output Quality
The quality lever the README now emphasises is visual review. In `skills/pptagent/config.yaml` you set `mode: text` or `mode: multimodal` and a `visual` block with `base_url`, `model`, `api_key_env` and `timeout_seconds`, and the key itself goes in `skills/pptagent/.env`. The README recommends `deepseek-v4-flash-vision` as the multimodal reviewer, with tokens available through the Duanyan Token Plan, and warns that the skill appends `/chat/completions` to `base_url`, so the value must be the API prefix from the console, not the Token Plan webpage URL. The 2026/01 news entry also notes offline mode support and context management added to prevent context overflow.
MCP Server and OpenClaw Integration
MCP server support was added in 2025/09, and the README points to the v1.1.37 documentation, `pptagent/DOC.md`, for details, which places it with the earlier runtime rather than the current Skill. On the current branch, MCP appears in the OpenCode path: the README sends OpenCode users to a separate setup guide in `skills/pptagent/README.md` for visual MCP configuration and the complete workflow.
Limitations and When It Is the Wrong Tool
The most obvious limitation is that Windows is unsupported, which excludes a large chunk of corporate users. Another is that the framework is agentic, meaning it makes autonomous decisions about slide structure and content. If you need a presentation that adheres to a strict corporate template with exact fonts, colors, and logo placement, PPTAgent is the wrong tool. The README mentions 'Freeform & template generation' in the January 2026 update, but that does not guarantee pixel-perfect fidelity to your brand guidelines. The output is a PPTX file that you will likely need to edit in PowerPoint or LibreOffice afterward. Also, the reflection stage may generate slides that are factually consistent with the source but stylistically bland.
Alternatives and How They Differ
The most direct alternative is to use a general-purpose LLM with a prompt like 'create a PowerPoint outline' and then manually copy the output into a slide tool. That approach gives you full control but no automation of the visual layout. Another alternative is to use a dedicated presentation tool like Gamma or Tome, which are cloud-based and offer template-driven generation. These tools are not open source and do not run locally, so they differ in privacy and customization. A more technical alternative is to use python-pptx directly to script slide generation, but that requires you to write code for every slide, which is the opposite of the agentic approach. PPTAgent sits between those extremes: it automates the entire pipeline but still produces an editable PPTX. The difference from Gamma is that PPTAgent runs through your own coding agent and produces an editable PPTX, while Gamma is a closed, template-based service. The difference from python-pptx is that PPTAgent uses an LLM to decide the content, not just the layout. For engineers who want a reproducible, scriptable pipeline, PPTAgent is closer to a CI/CD step than a design tool.
Maintenance and Upgrade Cost
The release list shows v0.2.0 on 2025-10-13, v2.0.0 ("PPTAgent V2: DeepPresenter Integrated") on 2025-12-16 and v1.1.38 on 2026-09-20, and the last push was on 2026-11-02, so the project is still active. The README directs anyone who wants the previous runtime or the papers' code to pinned versions rather than the current development branch: v0.2.0 for PPTAgent (EMNLP 2025) and v1.1.38 for DeepPresenter (ACL 2026). This is a fast-moving project, so expect the current branch to change. The project is licensed under MIT.
Editorial conclusion
Adopt PPTAgent if you already work in Claude Code, Codex CLI or OpenCode on Linux, WSL or macOS and want your coding agent to create, revise and visually review editable PowerPoint decks. Do not use it if you require fine-grained visual design or Windows-native support. Before adopting, verify that you have uv, npm and LibreOffice available as `libreoffice` on PATH, decide whether you want a multimodal visual reviewer and its API key, and use the pinned v0.2.0 or v1.1.38 tags if you need the earlier runtime or the papers' code.
Frequently asked questions
What is ppt agent, and what does PPTAgent add to the idea?
PPTAgent is an agentic framework for reflective PowerPoint generation, and the current branch ships it as a Skill for Claude Code, Codex CLI and OpenCode rather than as a standalone script. You register it from skills/pptagent/ and drive it through your coding agent.
Which PPTAgent tag reproduces the EMNLP 2025 paper?
v0.2.0, published 2025-10-13, is pinned to the EMNLP 2025 paper at arxiv.org/abs/2501.03936. The DeepPresenter paper accepted to ACL 2026 is pinned to v1.1.38 instead, published 2026-09-20.
Which coding agents can register the PPTAgent Skill?
Claude Code, Codex CLI and OpenCode. From skills/pptagent/ you run scripts/install.py with `--client claude`, `--client codex` or `--client opencode`, then run scripts/pptagent.py doctor to check the installation while the repository is still in place.
How is the visual reviewer configured in PPTAgent?
In skills/pptagent/config.yaml, set mode to text or multimodal and give the visual block a base_url, a model name, an api_key_env variable and timeout_seconds. The skill appends /chat/completions to base_url, so the value must be the API prefix from your console rather than the Token Plan webpage URL, and the key itself goes into skills/pptagent/.env.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/icip-cas-pptagent)Star history chart for your README
[](https://hysenlabs.com/projects/icip-cas-pptagent)Was this useful?