AI Long-Text Flow

English | 中文

English

Split long text with multiple Fragmentation strategies, then process each chunk with LLMs in Serial or Parallel execution. Designed for translation, polishing, analysis, and any workflow that runs into token limits or needs higher throughput.

Supports Google Gemini, OpenAI / Custom (any OpenAI-compatible API). API keys are stored locally in your browser; this project runs entirely on the client.

👉 Live Demo

Key Features

• Fragmentation (choose one):
  • Chars — split by fixed character count
  • Lines — split by fixed line count
  • Custom — split by Text Rule (wildcards / /regex/) or Headings (Markdown headings); optionally Keep marker
• Execution (choose one):
  • Parallel — process multiple chunks at once with a configurable Concurrency Limit; prompt injected per chunk
  • Serial — process chunks one by one; prompt injected per chunk
  • Serial + Contextual Memory — keeps conversation history across chunks; prompt injected only for the first chunk
• Payload preview — inspect the full JSON request (see the Payload tab)
• Real-time status — visual per-chunk state, retry failures, pause/stop anytime

More Features

• Merge into Input — merge completed request/result pairs back into the input for another round (great for iterative polishing)
• Split Rule Assistant — describe your split need in natural language and let AI generate the rule
• Glossary — bulk import terms and inject matched entries into the effective system prompt for consistent translation
• Chunks to Combine — combine small chunks to reduce API calls
• Export — Export Original Text / Export All Results / Export Combined (Source + Result), plus Custom Export

Quick Start

# 1. Clone & install dependencies
git clone https://github.com/zlhhhh8901/AI-Long-Text-Flow.git
cd AI-Long-Text-Flow && npm install

# 2. Dev
npm run dev

# Build for production
# npm run build

Acknowledgements

The fragmentation module’s UX and layout are inspired by ChatGPT Automator. Thanks to the original author for the inspiration.

中文

长文本多规则分块，串/并行调用 LLM 处理。解决长文本翻译、润色、分析等场景下的 Token 限制和效率问题。

支持 Gemini、OpenAI 及任意兼容 API。API Key 仅存浏览器本地，无泄露风险。

👉 在线体验

主要功能

• 多规则分块（任选其一）：
  • 按固定字符数分块
  • 按固定行数分块
  • 自定义 — 支持通配符 / 正则 / Markdown 标题拆分，可选保留或移除分隔符
• 三种执行模式（任选其一）：
  • 并行 — 多分块同时请求，可控并发数，每个分块独立注入提示词
  • 串行 — 逐块顺序请求，每个分块独立注入提示词
  • 串行 + 连续上下文 — 保持对话记忆，仅首个分块注入提示词，适合需要连贯性的场景
• 请求预览 — 执行工作流程前可检查完整 JSON 负载，方便调试
• 实时状态 — 可视化追踪分块状态，失败可单独重试，流程可随时暂停/终止

其他功能

• 往返编辑 — 合并请求与结果进行新一轮处理，适用于迭代润色等场景
• 文本分块助手 — 用自然语言描述需求，AI 生成分割规则
• 术语表 — 批量导入术语并自动注入提示词，确保翻译一致性
• 文本块合并 — 按规则合并小分块，减少 API 调用次数
• 灵活导出 — 纯结果 / 原文+结果对照 / 自定义格式

快速开始

# 1. 克隆并安装依赖
git clone https://github.com/zlhhhh8901/AI-Long-Text-Flow.git
cd AI-Long-Text-Flow && npm install

# 2. 开发模式
npm run dev

# 构建生产版本
# npm run build

致谢

文本分块模块的功能设计与界面布局参考 ChatGPT Automator，感谢原作者的启发。