Releases: ET06731/Memflow
Releases · ET06731/Memflow
v1.2.0 - Local AI Provider Support
New Features
- Add local AI Provider, support Ollama and LM Studio
- Auto-fetch model list in settings page
- Add test connection button for debugging
- Support OpenAI-compatible API for local models
How to Use
- Open extension settings page
- Select Provider as "Local AI"
- Fill in Base URL (Ollama: http://127.0.0.1:11434/v1, LM Studio: http://127.0.0.1:1234/v1)
- Select or enter model name
- API Key can be left empty
Known Issues
- Some local models may return empty response depending on model configuration
Memflow v1.1.5
v1.1.5
Fixes
- Improve YouTube subtitle extraction by waiting for Performance API timedtext URLs with
pottokens. - Fix YouTube timestamp links to preserve video IDs and generate valid
t=jumps. - Avoid React content-script render errors by converting the content script entry to a plain TypeScript script.
- Preserve YouTube/Bilibili embeds in exported Markdown.
Improvements
- Add English learning template support in popup and settings.
- Adjust popup template settings placement and save button position.
Package
- Chrome MV3 extension package attached:
Memflow-v1.1.5.zip.
v1.1.4
v1.1.3
v1.1.2
v1.1.1
v1.1.0
v1.0.6 release: 修复 Gemini 和豆包导出问题
v1.0.6 更新日志
- 修复 Gemini 导出问题:通过更精确的选择器和 BaseAdapter 逻辑改进,解决了多行输入被拆分和重复导出的问题。
- 修复 豆包 导出回归:修正了之前的提取逻辑错误,通过非空角色判定确保 Turn 容器正确拆分,避免回复合并到提问中。
- 通用提取引擎改进:
- 引入了基于纯文本的内容等效性判定。
- 支持相邻同角色消息片段的智能合并。
- 增强了 UI 噪声过滤。
v1.0.5 - 界面重构,设置增强,哔哩哔哩总结模板支持自定义
界面重构,设置增强,哔哩哔哩总结模板支持自定义。
v1.0.4 - B站视频智能总结提示词模板
新增功能:B站视频智能总结现在支持选择提示词模板(科技专栏/知识梳理)。