Tomato Novel Caps AI Writing After 104,000 June Rejections
Rest of World reported that Chinese web-novel platforms are adding limits after Tomato Novel said it rejected more than 104,000 low-quality submissions in June and Jinjiang restricted AI to research and proofreading.

Rest of World reported that China AI web novel platforms are moving from productivity push to quality control after writers used generative tools to mass-produce fiction for large reading apps.
The outlet reported that ByteDance, Tencent and Baidu own several of China's most influential web-novel platforms, where advertising, subscriptions and adaptation rights turn popular stories into a commercial pipeline.
ByteDance's Tomato Novel said it rejected more than 104,000 low-quality submissions in June, including AI-written work, after previously launching tools that helped authors research and continue stories.
Tomato Novel Rejected More Than 104,000 Low-Quality Submissions In June
Tomato Novel has capped how many words each account can publish per day.
Rest of World reported that the June rejection total included AI-written submissions, showing how a platform that benefits from high-volume writing is now limiting some of the same output.
Rest of World reported that Gordon Sheng, a 32-year-old civil engineer, used DeepSeek to outline a divorce plot and generated a short story in five minutes with another AI writing tool.
Rest of World reported that his story drew more than 5,500 reads in 10 days on Tomato Novel.
Sheng said AI helped him express a story he could not easily write alone, but he still had to correct mistakes.
One bug made a father in the story text his daughter even though the characters were standing next to each other.
China Literature Put AI Tools Around Qidian And Jinjiang
Rest of World reported that Tencent's China Literature introduced an AI writing tool in 2023 that could generate character names and expand a rough plotline into a fuller story.
China Literature chief executive Hou Xiaonan compared the shift at the time to moving from human driving to assisted driving.
China Literature owns Qidian and Jinjiang, two major Chinese web-novel platforms with different reader bases and business models.
Rest of World reported that the company also published more than 17,000 AI-translated stories on WebNovel, whose top markets include the U.S., India and Brazil.
Rest of World reported that ByteDance's Tomato Novel also offers an author's assistant for research and writing support.
The platform is free to read and earns from advertising, a model that gives it more tolerance for high-volume writing than subscription sites that need readers to pay for quality.
InkOS Downloads Passed 50,000 Users
The automation layer is becoming more elaborate than a single chatbot prompt.
Beijing developer Junxian Ma built InkOS, a fiction-writing system run by AI agents that assign different roles to an architect, writer and auditor while a human occasionally steers the plot.
Ma told Rest of World that more than 50,000 users had downloaded InkOS and that some users had earned money by publishing AI-generated stories on Tomato Novel.
The system shows why platforms are no longer only dealing with writers using AI as a drafting assistant; they are also dealing with agent systems built to produce entire serialised stories.
Reader resistance is part of the platform response.
Rest of World cited Xiaohongshu users sharing screenshots of prompts left inside novels, while reader Yang Zhou said fast-produced AI writing felt like a waste of time compared with slowly updated fiction.
Jinjiang Founder Limited AI To Research And Proofreading
Copyright risk has added another constraint.
Rest of World also reported that web-novel writers protested in 2024 after Tomato Novel required them to grant rights for AI training, and that the platform later allowed them to opt out of the programme.
Rest of World also reported that some AI tools promote features that extract a popular story's plotline and rewrite it as a new novel.
Jinjiang founder Huang Yanming told authors in a 2025 statement to use AI only for story research and proofreading.
Huang also asked readers to report suspected AI-written novels, while acknowledging that it remains difficult to determine how much of a story was produced with AI.
Xiang Ren, a University of Sydney senior lecturer who researches China's digital publishing industry, told Rest of World that the platforms are trying to balance AI's productivity gains with threats to authenticity.
Tomato Novel has not disclosed the daily word cap per account, how it detects AI-written fiction, or how many rejected submissions were written entirely by AI rather than partly assisted by it.


















