laihenyi/NBLM2PPTX：使用 Gemini AI 将 NotebookLM PDF

项目介绍

laihenyi/NBLM2PPTX 是一个托管在 GitHub 上的开源项目，主要使用 HTML 编写，采用 MIT License 许可证，目前获得 351 个 Star、76 次 Fork。

仓库全名: laihenyi/NBLM2PPTX

README 摘要

NBLM2PPTX - NotebookLM PDF to PPTX Converter Convert NotebookLM exported PDFs into PPTX presentations with separated background images and editable text layers. ✨ Updated (2026-10-21): NBLM2PPTX for macOS v0.1.1 — offline PDF → editable PPTX on your Mac, no API key, notarized by Apple. See “macOS Native App” below. 繁體中文 | 简体中文 | 日本語 | Español | Français macOS Native App (offline, Rust) NBLM2PPTX for macOS is a…

完整内容见 GitHub 原始 README 或本页。

常见问题

NBLM2PPTX 用什么语言开发？

主要使用 HTML 编写，采用 MIT License 许可证，目前有 351 个 Star。

NBLM2PPTX 可以免费商用吗？

如何获取 NBLM2PPTX 的源码？

可以直接克隆仓库：git clone https://github.com/laihenyi/NBLM2PPTX.git，默认分支为 main；安装与运行步骤见仓库 README。

NBLM2PPTX 最近还在维护吗？

仓库最近一次更新为 4 天前，当前没有未关闭的 Issue；Star 数为 351，可作为社区活跃度的参考。

关于

GitHub 原始描述（英文）

Convert NotebookLM PDFs to PPTX with separated background images and editable text layers using Gemini AI

ai-tool gemini-ai google-gemini notebooklm ocr pdf-converter pptx text-removal pdf-to-pptx

MIT License
351 Stars
76 Forks
3 Watching

描述取自 GitHub 官方 API，保持仓库作者书写的原文。

简要分析

热度分

506

价值分

233

活跃状态

活跃

GitHub

语言占比

HTML

Python

Rust

Shell

TypeScript

Release

NBLM2PPTX for macOS 0.1.1（原生離線版）3 周前
NBLM2PPTX for macOS 0.1.0（原生離線版）3 周前
v2.3 Multilingual Demo Videos with Background Music1月22日
v2.3 Multilingual Demo Videos1月22日
v2.3 Demo Video1月22日

贡献者

laihenyi

NBLM2PPTX - NotebookLM PDF to PPTX Converter

Convert NotebookLM exported PDFs into PPTX presentations with separated background images and editable text layers.

✨ Updated (2026-10-21): NBLM2PPTX for macOS v0.1.1 — offline PDF → editable PPTX on your Mac, no API key, notarized by Apple. See “macOS Native App” below.

繁體中文 | 简体中文 | 日本語 | Español | Français

macOS Native App (offline, Rust)

NBLM2PPTX for macOS is a native app that runs entirely on your Mac — no API key, nothing uploaded. Source code: macos-app/.

Download: release page (NotebookLM2PPT-notarized.zip)
OCR: macOS built-in Vision (on-device)
Text removal: erases only the text strokes; textured backgrounds are rebuilt by LaMa running on the GPU (CoreML)
Object extraction (experimental): icons and illustrations become movable images (MobileSAM)
App: drag-and-drop GUI, signed with Developer ID and notarized by Apple; models (~253 MB) are downloaded on first launch
Requirements: Apple Silicon; macOS 12+ for GPU inference

Build, CLI usage and design notes: macos-app/README.md. Third-party licenses: macos-app/THIRD-PARTY.md.

Before / After (standard version)

Original (NotebookLM PDF)	Output (editable PPTX, macOS app)

Text becomes editable text boxes; illustrations stay in the background and the NotebookLM watermark is removed (output rendered with LibreOffice). Current limitations visible here: highlight colors are not kept, bold is sometimes lost, and OCR may misread a few characters.

v0.1.1 (2026-10-21)

Object version about 2× faster, peak memory ~6 GB → ~1.9 GB: SAM now runs on the CPU only (11-page deck: 47–86 s, was 103–183 s)
Multi-line body text is left-aligned when the original is (it used to be always centered)
Faster mask dilation for object extraction
GUI: the model section shows whether the GPU version of LaMa is ready
Developer tooling: the PowerPoint compatibility check only reads and closes the presentation it opened itself

v0.1.0 (2026-10-21)

First release: drag-and-drop app, signed and notarized
Speed (Apple Silicon, 11-page deck): standard version 29–45 s (~1.1–1.4 GB), object version 103–183 s (~6 GB)

Gemini-Assisted Web Version (no longer updated)

The earlier Gemini-assisted web version (index.html and its language versions, demo video project) is archived in legacy-web/. It is no longer updated; developers who need it can take it from there. Its documentation: legacy-web/README.md.

License

MIT License

暂无评论

NBLM2PPTX