Open Notebook 1.1.1 — macOS Desktop

Bump version to 1.1.1 across pyproject.toml, uv.lock, and npm packages.

Desktop:

• Add build_desktop_release.sh for .app + .dmg packaging
• Add GitHub Actions workflow (desktop-release.yml)
• Translate desktop/README.md to English and document CI

Documentation:

• Rewrite root README for eryabyshev/open-notebook-app fork

Fixes and improvements (included in this release):

• Transformations output in source language
• Fix embedding command registration on sync /api/embed path
• GitHub version check configured for fork repository
• Sign out quits the desktop app
• API keys badge contrast in light theme
• Source detail modal header action overlap

Open Notebook Desktop v0.1.0 — macOS (Apple Silicon)

Open Notebook Desktop v0.1.0

First public desktop app release for macOS (Apple Silicon).
Everything runs locally: no Docker required — SurrealDB, API, worker, and UI in a single installer.

Install

1. Download Open Notebook-0.1.0-arm64.dmg
2. Open the DMG and drag Open Notebook into Applications
3. Launch the app
4. Go to Settings → API Keys, add your provider (OpenAI, OpenRouter, etc.), and set default models

What's included

• Full UI (Next.js standalone)
• Local SurrealDB v2 database
• Background worker (embeddings, source processing, podcasts)
• 14 UI languages
• Update checks via GitHub (eryabyshev/open-notebook-app)

Highlights

• Upload sources (PDF, URLs, audio, text)
• Chat with citations, semantic search, transformations
• Insights and summaries in the source language
• Podcasts, model configuration via Settings
• Optional Docling extension for advanced PDF processing (download from Settings)

Requirements

• macOS, Apple Silicon (arm64)
• ~2 GB free disk space (lite build, Docling not included)
• Internet connection for AI providers (OpenAI, etc.)

Docling (optional)

Advanced PDF/OCR processing is not bundled in the DMG.
Install via Settings → Advanced PDF processing (Docling) → Download from GitHub.

Known limitations

• Build targets arm64 only (Intel Mac requires a separate darwin-amd64 build)
• macOS may warn about an unsigned app on first launch
• Embedding model must be configured in Settings for vector search

For developers

• Source: https://github.com/eryabyshev/open-notebook-app
• Build: bash desktop/build_packaged_app.sh
• Docs: docs/0-START-HERE/quick-start-openai.md

---

Tag: v0.1.0
Asset: Open Notebook-0.1.0-arm64.dmg

Docling Extension v1.1.1 — fix Pillow/PyMuPDF native libs (macOS arm64)

Docling Extension v1.1.1

Optional desktop extension for Open Notebook: advanced PDF processing (Docling + OCR on macOS via ocrmac).

What's fixed

• Pillow / libtiff — materialized dylibs survive install from tar.gz (no broken Homebrew symlinks)
• PyMuPDF / libmupdfcpp.so — bundled and materialized at _internal/ root; graphs load correctly after install
• Extension install no longer fails with "missing graphs / Pillow native libraries"

Assets (per platform)

Each platform ships two archives (GitHub 2 GiB limit):

File	Contents
docling-extension-worker-{platform}-{version}.tar.gz	Docling PyInstaller worker (~1.5 GB)
docling-extension-models-{platform}-{version}.tar.gz	Docling OCR/layout models (~1.1 GB)
docling-extension-{platform}-{version}.manifest.json	Manifest for the desktop app
*.sha256	Checksums

Platforms: darwin-arm64, darwin-x64, linux-x64, windows-x64 (via CI matrix on publish)

Install in Open Notebook Desktop

1. Settings → Advanced PDF processing (Docling)
2. Download from GitHub — or Install local file… (select both worker + models .tar.gz)
3. Quit app completely (Cmd+Q) → reopen
4. Upload a scan PDF — worker.log should show Docling artifacts: …/models and document_engine='docling'

Upgrade from v1.1.0-extension

1. Remove old extension (Settings or bash desktop/clean_docling_extension.sh --yes)
2. Install v1.1.1 worker + models from this release
3. Restart app

Optional one-liner if libs break after manual copy: bash desktop/repair_docling_worker_dylibs.sh

Requirements

• Open Notebook Desktop with lite worker (default installer)
• ~4 GB free disk space for extension (worker + models)

Build

Built with bash desktop/build_worker_docling.sh — native libs verified before packaging.