开源免费识屏搜索工具 eSearch 中文版由大眼仔旭（www.dayanzai.me）发布。文本识别软件现在有点奇怪。虽然一些用户可能会看到他们的潜力，但他们的实用性充其量是值得怀疑的。更尖锐的是，这种解决方案在哪种情况下可行？这也与软件识别文本的效率有多高有关。

eSearch 就是这样一个屏幕识别程序，它旨在为用户提供一个简单易用、高效且通用的解决方案，以满足多种需求。查找文本和图像、截取屏幕截图、记录屏幕区域、在显示屏上注释元素等等。

eSearch 功能特色

截屏
框选裁切
框大小位置可调整(支持方向键或 WASD)
取色器
放大镜
画笔（自由画笔）
几何形状（边框填充支持调节）
高级画板设置（使用 Fabric.js 的 api）
图像滤镜（支持局部马赛克模糊和色彩调节）
自定义框选松开后的操作
快速截取全屏到剪贴板或自定义的目录
截图历史记录
保存（可选保存为 SVG 可编辑文件）
其他应用打开
复制到剪贴板
钉在屏幕上
滚轮缩放
恢复默认大小位置
透明度
鼠标穿透
二维码识别
OCR 识别
离线 OCR（基于 C++ 或 Python 实现，eSearch-OCR）
自定义离线 OCR 模型和字典
其他在线 OCR
在线公式识别
支持自己申请秘钥
以图搜图
托盘
划词句来搜索
识别展示
自动搜索翻译
搜索
翻译
自定义搜索翻译引擎
软件自带浏览器打开
跟随关闭、失焦关闭
系统浏览器打开
链接识别
历史记录
自动删除换行（用于自动排版）
查找替换（支持正则匹配）
其他编辑器编辑（支持自动重载）
行号
拼写检查

eSearch 是一个雄心勃勃的程序，但有时感觉效率很低。尽管如此，它还是值得关注的，因为潜力是存在的。

官方主页

What’s new in eSearch 1.12.1
Jul 8, 2024
Program:
Added new main page mode to force the main page to search or translate
Fix some systems without cli
Fix Linux appimage slow startup
Fix rpm dependency error #241
The optimized main page and settings cannot be jumped directly. To open the settings, you need to open the settings in the tray or menu bar.
Screenshot:
Added multi-screen switcher
Added free screenshots, you can capture any shape #94
Added the ability to display the toolbar later to prevent it from being blocked when taking screenshots, and supports shortcut keys to show or hide #219
Added tray OCR and image search to temporary “default operation after frame selection” and supports shortcut key binding.
Added the ability to save and copy settings
Added WebP format
Optimize to ensure that the window is displayed after taking a screenshot to prevent the screenshot window from displaying animations
Optimize the removal of OCR and image search scan lines, and display the main page directly
Optimize the QR code recognition operation to be performed on the main page
Optimize Optimize color picker performance
The optimized size bar can be adjusted with the direction keys, supports Ctrl and Shift key acceleration, can be focused with shortcut keys, and can be used as a new operation for shortcut frame selection.
Optimize OCR selection menu style
Optimize esc exit brush mode and other logic
Fix Fixed cursor direction error
Edit (draw):
Added the function of clicking the edit bar main menu and automatically selecting the submenu.
Added support for shortcut keys in some main menus and submenus
Added alt display shortcut key
Added arrow keys and scroll wheel to set border width and brush thickness
Newly added controls such as border width and filter effect size can be adjusted using the up and down keys.
Added spotlight shape to quickly create hollow graphics like frame selection
Added numbers and arrows to record color attributes
Optimize the main menu suspended in the edit bar to display the submenu
There is no need to select the filter first to optimize the filter. You can select the filter and then select the frame.
Optimized for graphics such as lines, brushes, arrows, etc., the edit bar property preview only displays strokes, and the color wheel defaults to the stroke color wheel.
Optimize and adjust the stroke value range
Fixed the problem that the color wheel cannot be selected as grayscale. Now you can select black and white in the color wheel.
Long screenshot (scrolling screenshot):
Added preview screenshot
Added universal splicing
Optimize real-time stitching
Optimize the end button to prevent occlusion
Screen translation:
New features. Translations can be generated near box selections
Manual mode can directly cover the frame selection area and translate the text on the screen. It is suitable for temporary translation and picture translation.
Automatic mode, above or below the frame selection area, regularly detects screen updates, and then displays the translation, suitable for game translation
Supports adding translators in settings, supporting Baidu, Youdao, Bing, Caiyun, deepl, chatgpt, gemini, Mavericks, Tencent Interactive, Tencent, Huoshan, see xtranslator for details
Ollama (deploying local large models) is compatible with ChatGPT api. Please note that you can fill in the key in the settings casually.
OCR:
Added Baidu OCR form recognition
Optimized model updated to paddleOCR v4
AI:
Added new backends that can set local OCR and screen recording virtual background, such as CPU, CUDA (NVIDIA GPU), CoreML (Apple acceleration), etc.
Record screen:
Added custom size and position of key prompts
Added extension of key prompt display time
Added key prompts to support merging shortcut keys #155
Added selection of cameras and audio devices
Fix log display error
Fix scaling bug
Fix cannot be saved
Main page editor:
Added export table to csv format (can be edited and saved as in excel) and markdown table
Added home page shortcut keys
New reuse window, OCR and other results are not displayed in a new window, but the previously recognized pop-up window is reused, which can save time for frequent OCR tasks
Newly added advanced window buttons – top, auto-close when focus is lost, and simple mode buttons
Add a new window to the top, like a sticker, for easy comparison
The new function of automatically closing when losing focus is extracted from the settings, which can be more conveniently controlled and specify which windows are automatically closed and which windows are retained.
A new simple mode is added, and the interface only retains text editing. Search and other operations are performed through shortcut keys or selected text
Optimize the text size of the picture area to adapt to the original text size of the picture
Optimize and replace the “translation tree” with a local interface, sharing the API with screen translation
Fix picture area selection error
Fixed an issue where part of the text was truncated after enlarging the font
Fix toolbar not following
Stickers:
Added the ability to set shadow prompts to always be displayed
fix minimize
Fix unable to edit
Search pictures by pictures:
Fix google image search error
Set up:
Optimized search translation engine change method updated, more intuitive adjustment
Fix write errors
Develop:
Multiple languages can be exported to csv for editing
Developer mode can be turned on in settings

资源：629.rar
解压密码：www.dayanzai.me
转载请保留出处，谢谢合作~
点击下载（提取码：332y）
点击下载（提取码：8fi5）
点击下载（提取码：536766）
点击下载（提取码：536766）