.
\ No newline at end of file
diff --git a/README.md b/README.md
index 3bd03cf..2350d73 100644
--- a/README.md
+++ b/README.md
@@ -2,36 +2,38 @@
OpenRewind is an open-source alternative to [rewind.ai](https://rewind.ai), forked from [OpenRecall](https://github.com/openrecall/openrecall).
-As this project is still in its early stages, you might not see significant progress just yet.
+We wanted to create an open source app that provides similar core functionality
+to rewind.ai, and that app is **OpenRewind**.
## To-dos
### Update the OCR Engine
-OpenRecall currently uses docTR as its OCR engine, but it performs inadequately. On my MacBook Air M2 (2022), processing a screenshot takes around 20 seconds, with CPU usage peaking at over 400%. During this time, screenshots cannot be captured, and the engine appears to recognize only Latin characters.
+OpenRecall currently uses docTR as its OCR engine, but it performs inadequately.
+On my MacBook Air M2 (2022), processing a screenshot takes around 20 seconds, with CPU usage peaking at over 400%.
+During this time, screenshots cannot be captured, and the engine appears to recognize only Latin characters.
-To address this, we plan to replace the OCR engine with a more efficient alternative that supports multiple writing systems. We are considering [Tesseract](https://github.com/tesseract-ocr/tesseract), which supports multiple languages and is faster, though it still falls short of our desired results. (As reference, rewind.ai runs on same machine with avearge CPU usage of 40%)
+To address this, we plan to replace the OCR with a more efficient alternative that supports multiple writing systems.
+We are working on [RapidOCR ONNX](https://github.com/alikia2x/RapidOCR-ONNX), a fork of a project which has same name,
+developed by RapidAI.
+RapidOCR ONNX uses [PaddleOCR](https://github.com/PaddlePaddle/PaddleOCR) as its model architecture, and
+runs on the [ONNX Runtime](https://github.com/microsoft/onnxruntime/).
### Implement a Task Queue/Scheduler
-Currently, OpenRecall's OCR recognition and database operations are synchronous (blocking). This results in increased screenshot frequency, as described in the previous section.
+Currently, OpenRecall's OCR recognition and database operations are synchronous (blocking).
+This results in increased screenshot frequency, as described in the previous section.
Our next goal is to introduce a task queue to handle high-load tasks (such as OCR, indexing, and archiving) asynchronously. This will ensure that time-sensitive tasks (like capturing screenshots) are prioritized.
### Improve the Frontend
-The current frontend of OpenRecall is quite basic. Given my expertise in web development, I will build a more robust frontend from scratch, using Python solely as a daemon/backend server.
+The current frontend of OpenRecall is quite basic. Given my expertise in web development,
+I will build a more elegant frontend from scratch.
-We are also considering using Electron to deliver a near-native experience, aiming to match the functionality of [rewind.ai](https://rewind.ai).
+We are now switched to Electron in order to deliver a native experience,
+aiming to match the functionality of [rewind.ai](https://rewind.ai).
### Add More Features
We will be implementing the [feature list](https://github.com/openrecall/openrecall/discussions/9) proposed in the OpenRecall repository. Stay tuned for updates.
-
-### THE FURTHER FUTURE: Deployment & Refactoring
-
-We are exploring ways to simplify the download, installation, and usage of this Python-based program, especially as we introduce new technologies (such as web dev / Electron).
-
-Initially, we are considering Docker as a deployment solution. [Docker Desktop](https://www.docker.com/products/docker-desktop/) is available for macOS and Windows users and offers a more user-friendly experience compared to command-line Python installations and dependency management.
-
-However, Docker is a temporary solution. In the future, we plan to transition OpenRewind to the [Electron](https://www.electronjs.org/) technology stack and gradually migrate Python components to Node.js.
diff --git a/components/settings/OpenSourceNote.tsx b/components/settings/OpenSourceNote.tsx
index 39daa37..d03c2e4 100644
--- a/components/settings/OpenSourceNote.tsx
+++ b/components/settings/OpenSourceNote.tsx
@@ -5,7 +5,7 @@ export default function OpenSourceNote() {
OpenRewind is open source software licensed under
- AGPL 3.0.
+ GPL 3.0.
Source code is avaliable at
GitHub
diff --git a/i18n/en.json b/i18n/en.json
index 5022ed3..60fae98 100755
--- a/i18n/en.json
+++ b/i18n/en.json
@@ -6,7 +6,7 @@
"screen-recording": "Screen Recording",
"version": "Version {version}",
"copyright": "Copyright © 2024 alikia2x",
- "note": "OpenRewind is an open source software licensed under <1>AGPL 3.01>, <3>3>and its source code is avaliable at <5>GitHub5>.",
+ "note": "OpenRewind is an open source software licensed under <1>GPL 3.01>, <3>3>and its source code is avaliable at <5>GitHub5>.",
"environment-details": "Environment Info",
"node-version": "Node.js Version",
"electron-version": "Electron Version",
diff --git a/i18n/zh-CN.json b/i18n/zh-CN.json
index 70a9ca2..98ca4a9 100644
--- a/i18n/zh-CN.json
+++ b/i18n/zh-CN.json
@@ -6,7 +6,7 @@
"screen-recording": "屏幕录制",
"version": "版本 {version}",
"copyright": "版权所有 © 2024 alikia2x",
- "note": "OpenRewind 是基于<1>AGPL 3.01>授权的开源软件, <3>3>其源代码托管在<5>GitHub5>上。",
+ "note": "OpenRewind 是基于<1>GPL 3.01>授权的开源软件, <3>3>其源代码托管在<5>GitHub5>上。",
"environment-details": "运行环境",
"node-version": "Node.js 版本",
"electron-version": "Electron 版本",