#3939th of 60 in Desktop AI apps

pagekin

Desktop app for writing novels chapter by chapter with your own LLM

Stars
2.7k
License
GPL-3.0
Last commit
Nov 2026
Last release
Nov 2026
Language
TypeScript
Stars, last 21 weeks+2,639
20 Jun 2026: 46 stars27 Jun 2026: 46 stars4 Jul 2026: 47 stars11 Jul 2026: 47 stars18 Jul 2026: 140 stars25 Jul 2026: 270 stars1 Aug 2026: 331 stars8 Aug 2026: 365 stars15 Aug 2026: 399 stars22 Aug 2026: 470 stars29 Aug 2026: 525 stars5 Sept 2026: 684 stars12 Sept 2026: 853 stars19 Sept 2026: 1,041 stars26 Sept 2026: 1,156 stars3 Oct 2026: 1,304 stars10 Oct 2026: 1,416 stars17 Oct 2026: 1,619 stars24 Oct 2026: 1,828 stars31 Oct 2026: 2,209 stars7 Nov 2026: 2,481 stars14 Nov 2026: 2,685 stars

46 on 20 Jun 20262,685 on 14 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Pagekin (formerly AI Novel Writer) is a Windows and macOS desktop app that takes a novel through settings, architecture, chapter blueprints, drafting, review, revision and finalization. Works are stored locally in SQLite, and generation uses your own OpenAI-compatible or Gemini API key, or a local Ollama model. It can also import .txt, .md or .epub novels for style analysis or continuation, and load prompt-only SKILL.md writing rules.

Who it is for: Fiction authors writing in Chinese or English who want a local workflow

Strengths

  • Works stored in a local folder plus SQLite database
  • Revisions shown as per-paragraph diffs instead of overwriting the draft
  • Each chapter prompt uses blueprint, characters, worldbuilding and summaries, not full chat history
  • Presets for DeepSeek, OpenAI, Gemini, xAI and Ollama via OpenAI-compatible /v1

Weaknesses

  • No bundled model; needs your own API key or local Ollama
  • Only OpenAI-compatible and Gemini protocols; Anthropic and Azure need separate adapters
  • Windows installer unsigned; macOS builds not notarized and no auto-update there
  • Assistant chat hidden in 1.3.0; batch writing capped at 10 chapters per run

What it needs

  • GPU optional
  • Needs LLM API key or Ollama, Embedding model (optional, for semantic search), Node.js 22.14+ and pnpm (source build only)
  • Models: OpenAI-compatible Chat Completions, Google Gemini, DeepSeek, SiliconFlow, BigModel

What changed

  1. v1.4.0 ↗ (opens in a new tab)Security fix

    “The runtime is upgraded to Electron 44, which includes upstream security fixes. File dialogs open in the folder you used last.”

From the release notes. No flag means the note does not say so, not that the upgrade is safe.

Also in Desktop AI apps

See all 60
Also in Desktop AI apps
RankProjectScore
1cherry-studioDesktop client for chatting with many LLM providers, with agents and knowledge bases53.6k starsAGPL-3.092 out of 100
2dsh-desktopNative Windows and macOS desktop client for DeepSeek Harness, built on plugins36.3k starsMIT91 out of 100
3screenpipeRecords screen and audio locally, makes it searchable for you and your agents22.4k starscustom license91 out of 100
#4hermes-desktopDesktop GUI for installing and chatting with Hermes Agent14.6k starsMIT87 out of 100
#5prompt-optimizerPrompt rewriting and testing tool for web, desktop, Chrome and MCP Live demo ↗ (opens in a new tab)41.9k starscustom license86 out of 100