<div align="center" id="中文">
<p align="center">
<img src="assets/logo.png" alt="EasySlides Logo" width="128" height="128">
</p>
<h1 align="center">EasySlides</h1>
<p align="center">
<a href="CHANGELOG.md"><img alt="Version: 1.2.1" src="https://img.shields.io/badge/Version-v1.2.1-2563EB?style=flat-square"></a>
<a href="LICENSE"><img alt="License: MIT" src="https://img.shields.io/badge/License-MIT-16A34A?style=flat-square"></a>
<img alt="Output: Editable PPTX" src="https://img.shields.io/badge/Output-Editable%20PPTX-D24726?style=flat-square">
</p>
面向论文汇报、研究综述与答辩的 AI PPT 插件,让 Agent 基于研究材料制作可编辑 PPT。
继续开发前请先读 [项目交接入口](docs/agent-handoff.md),再按其中的工作流与验证入口执行。
[中文](#中文) | [English](#english)
</div>
### 先看效果
| 原生可编辑 · `literature_minimal` | 图片整页重建 · `zdyf_blue` |
| --- | --- |
| [](http://easyslides.scansci.com/?view=detail&work=0) | [](http://easyslides.scansci.com/?view=detail&work=5) |
点击缩略图查看整套 PPT;展厅支持逐页预览与下载可编辑文件。
[成品与原生模板](http://easyslides.scansci.com/?library=native#catalog) · [图片参考模板](http://easyslides.scansci.com/?library=image#catalog)
原生模板用于直接制作,图片模板用于指导生图。也支持提供自己的 PPT 或参考图片。
### 安装与开始
在 Codex 或支持技能安装的 AI Agent 中发送:
```text
请帮我安装 EasySlides 的完整仓库和运行依赖,并检查能否生成 PPT:
https://github.com/Rimagination/easyslides
```
安装后,上传材料并说出目标:
```text
用 easyslides 把这篇论文做成组会汇报,先逐项弹窗问清需求,再开始制作。
```
插件会询问尚未明确的材料来源、听众、目标、时长、风格与制作路径,说明相对耗时和
Token 消耗;已知信息不重复问,无需浏览器或填表。安装细节见 [安装指南](INSTALL.md)。
### 三条制作路径

复制下面一句,替换主题、材料或篇幅即可。
#### 1. 直接制作:便于反复修改内容与布局
文字与布局直接构建为可编辑元素,论文原图作为图片保留。
```text
用 easyslides 基于我上传的多模态城市交通论文,
使用 literature_minimal 模板直接制作 12 页可编辑综述,
保留原图和来源,不补充外部资料。
```
#### 2. 整页重建:先确定完整图片页,再忠实重建
以下案例重建整页文字与简单结构,复杂配图保留为图片;已有完整图片页时可跳过生图。
```text
先用联网工具搜集多模态城市交通流量预测的论文与原图,再用 easyslides 制作 12 页综述:
每次 ImageGen 附上“国自然-绿色”模板参考图和适用的论文原图,逐页生成完整图片 PPT,
再忠实整页重建为可编辑 PPT,保留复杂配图。
```
#### 3. 局部重建:只让指定区域可编辑
选中区域内的文字与简单结构可编辑,其余部分保持图片。
```text
用 easyslides 局部重建我上传的多模态城市交通汇报图片,
只让标题、正文和页码可编辑,保留复杂配图;先确认重建区域,其余部分保持原貌。
```
### 使用须知
- **工具依赖**:联网搜集依赖宿主工具,EasySlides 未内置 Deep Research 引擎;新图片生成需要可调用的 ImageGen。
- **质量检查**:图片重建先确认样页,再扩展整套;交付前核对原文证据、页码和实际渲染。详见 [重建工作流](workflows/slide-image-to-editable-pptx.md)。
- **隐私与授权**:文件在本地管理,提交给云端 AI 的材料由对应服务处理;发布私人材料及复用第三方图片前需确认授权。
<details>
<summary><strong>开发与致谢</strong></summary>
开发文档:[架构说明](ARCHITECTURE.md) · [工作流入口](workflows/index.md)。
贡献者名单见 [CONTRIBUTORS.md](CONTRIBUTORS.md)。
EasySlides 的开发借鉴了以下公开项目的实践:
- 可编辑 PPTX 与本地生成:[hugohe3/ppt-master](https://github.com/hugohe3/ppt-master)、[Gabberflast/academic-pptx-skill](https://github.com/Gabberflast/academic-pptx-skill)。
- 学术表达:[LearnPrompt/humanize-ppt](https://github.com/LearnPrompt/humanize-ppt)。
- 风格与模板管理:[op7418/guizang-ppt-skill](https://github.com/op7418/guizang-ppt-skill)。
- 论文与文献报告:[xiao634zhang/paper-ppt-skill](https://github.com/xiao634zhang/paper-ppt-skill)、[fangyuanopus/literature-report-ppt-builder](https://github.com/fangyuanopus/literature-report-ppt-builder)。
致谢不代表这些项目对 EasySlides 的正式背书。代码许可见 [LICENSE](LICENSE);引用、上传或再分发第三方模板与图片时,请确认对应素材的授权。
</details>
[回到顶部](#easyslides)
---
## English
A research-focused AI PPT plugin for paper presentations, literature reviews and thesis defenses.
### See the Results
[Examples and native templates](http://easyslides.scansci.com/?library=native#catalog) · [Image-reference templates](http://easyslides.scansci.com/?library=image#catalog)
Native templates support direct creation; image references guide image generation. You can also supply your own PPTX or reference images.
### Install and Start
In Codex or an Agent that supports skill installation, say:
```text
Please install the full EasySlides repository and its runtime dependencies,
and check that it can generate PPTX files:
https://github.com/Rimagination/easyslides
```
Then ask:
```text
Use easyslides to turn this paper into a journal-club presentation.
Ask me about the requirements step by step before creating it.
```
EasySlides asks about missing source, audience, goal, timing, style and production choices, with relative time and token estimates. Known answers are reused; no browser form is required. See the [installation guide](INSTALL.md) for the full repository and dependencies.
### Three Methods, Three Examples
#### 1. Direct creation — revise content and layouts freely
```text
Use easyslides to create a 12-slide editable review from my uploaded
multimodal traffic papers with literature_minimal.
Preserve original figures and citations; do not add external sources.
```
#### 2. Full-slide reconstruction — establish the image, then rebuild it
```text
First use web tools to collect multimodal traffic prediction papers and original figures.
Then use easyslides to make a 12-slide review: attach the “国自然-绿色” template references
and relevant paper figures to each ImageGen call, generate complete slide images,
and faithfully rebuild them with editable text and simple structures,
preserving complex pictures.
```
Approved slide images can also be supplied directly, without new image generation.
#### 3. Partial reconstruction — edit selected regions only
```text
Use easyslides to rebuild only the titles, body text and page numbers
in my traffic presentation images. Confirm the regions first;
preserve complex pictures and leave everything else unchanged.
```
### Before You Use It
- **Dependencies:** web research uses host tools; EasySlides has no built-in Deep Research engine. New image generation requires a callable ImageGen tool.
- **Quality:** review representative rebuilt slides before expanding the deck, then check evidence, pagination and actual renders. Preserved pictures and unselected regions remain images. See the [reconstruction guide](workflows/slide-image-to-editable-pptx.md).
- **Privacy:** files stay in your workspace; cloud AI processes submitted material. Confirm permission before publishing private files or reusing third-party images.
<details>
<summary><strong>Implementation and Credits</strong></summary>
See [ARCHITECTURE.md](ARCHITECTURE.md) and [INSTALL.md](INSTALL.md) for development details.
Contributors are listed in [CONTRIBUTORS.md](CONTRIBUTORS.md).
EasySlides draws on public practice from [hugohe3/ppt-master](https://github.com/hugohe3/ppt-master), [Gabberflast/academic-pptx-skill](https://github.com/Gabberflast/academic-pptx-skill), [LearnPrompt/humanize-ppt](https://github.com/LearnPrompt/humanize-ppt), [op7418/guizang-ppt-skill](https://github.com/op7418/guizang-ppt-skill), [xiao634zhang/paper-ppt-skill](https://github.com/xiao634zhang/paper-ppt-skill), and [fangyuanopus/literature-report-ppt-builder](https://github.com/fangyuanopus/literature-report-ppt-builder). Acknowledgement does not imply endorsement. See [LICENSE](LICENSE) for the code license; third-party templates and images require their own permissions.
</details>
[Back to top](#easyslides)