Generate chapter-by-chapter summaries of very long web novels using a local or hosted AI model.
Track characters, events, and plot threads across thousands of chapters in a knowledge graph.
Resume or reprocess chapter summaries as new chapters of a story are released.
| daviid-p/ai-summary | 0xustaz/streamgate | a-bissell/unleash-lite | |
|---|---|---|---|
| Stars | 1 | 1 | 1 |
| Language | Python | Python | Python |
| Setup difficulty | moderate | hard | hard |
| Complexity | 3/5 | 4/5 | 4/5 |
| Audience | developer | developer | researcher |
Figures from each repo's GitHub metadata at analysis time.
Requires Ollama or an OpenAI-compatible API running locally or remotely, plus a folder of chapter text files.
This project is a Python script that reads long web novels, chapter by chapter, and generates a plain-English summary for each one using a large language model, the same kind of AI technology behind chatbots. It was written specifically to handle very long stories, thousands of chapters in length, by keeping track of characters, events, and story details across the whole novel rather than summarizing each chapter in isolation. To do this, the tool builds and updates what it calls a knowledge graph as it works through the chapters, a structured record of characters, their traits, ongoing plot threads, and important events. As the story grows, this record can get too large to hand entirely to the AI model each time, so the tool periodically compacts it, trimming or archiving details about characters and events that have not been relevant recently, while keeping full records of everything in a separate archive file just in case. The script is designed to run with Ollama, a tool for running AI language models directly on a person's own computer rather than through an online service, though it can also connect to OpenAI compatible services if a user prefers a hosted model instead. Someone using it points the script at a folder of chapter files, and it processes them one at a time or in small batches, writing a short summary for each chapter along with an ongoing log of story events. It can resume where it left off if interrupted, skip chapters already summarized, or be told to start over from scratch. The output includes individual chapter summaries, the running knowledge graph in its current and archived forms, an event log tracking what happened and who was involved, and a processing log recording how the run went. The author is upfront that the entire script was written by an AI assistant and shared as is because it worked for their own specific use case. They encourage anyone trying it to read through the code and test it carefully rather than trusting it without checking, and they welcome bug reports and contributions from others.
A Python tool that uses a local AI model to summarize long web novels chapter by chapter, while tracking characters and plot details in an ongoing knowledge graph.
Mainly Python. The stack also includes Python, Ollama, OpenAI API.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.