OCR Text Recognition Services
OCR extracts text from images, screenshots, and scanned materials. After recognition, you can continue with document summarization, AI chat, note organization, and knowledge base analysis.
WiseMindAI supports both local recognition and online OCR services. On macOS, you can use Apple OCR directly. On other systems, you can use an Ollama vision model or configure online services such as Baidu OCR, Baidu Qianfan, and Tencent Cloud OCR.
Fees for online OCR are charged entirely by the model provider and are not related to WiseMindAI.
Supported OCR services
| Provider | Application | Notes |
|---|---|---|
| Apple OCR | No application required | macOS only; uses the system's built-in text recognition |
| Ollama | No application required | Requires a vision-capable model installed locally |
| Baidu OCR | Open Console | Requires an API Key and Secret Key |
| Baidu Qianfan OCR | Get API Key | Supports models such as PaddleOCR-VL and DeepSeek-OCR |
| Tencent Cloud OCR | Open Console | Requires a SecretId and SecretKey |
1. Use Apple OCR
Apple OCR is only shown on macOS and does not require a key. Open Settings → Model Settings → Image Text Recognition, select Apple OCR, and set it as the active model.
Apple OCR uses the Mac's built-in text recognition. It is a good choice when you want to reduce third-party calls or quickly process screenshots and ordinary images.
2. Configure Baidu OCR
Open the Baidu OCR console. Create an account if necessary, then follow these steps:
- Go to Public Cloud Services → Application List and create an application.
- Return to the application list and copy its API Key and Secret Key.
- Open Settings → Model Settings → Image Text Recognition in WiseMindAI.
- Select Baidu OCR, then enter the API Key, Secret Key, and recognition version.
- Save and select Test.
3. Configure Baidu Qianfan OCR
Baidu Qianfan OCR uses a Qianfan API key and supports models such as PaddleOCR-VL, DeepSeek-OCR, and PP-StructureV3.
- Open the Baidu Qianfan API Key page and create or copy a key.
- Select Baidu Qianfan under Image Text Recognition in WiseMindAI.
- Enter the API key, choose an OCR model, and confirm the endpoint.
- Save and test the recognition result.
4. Configure Tencent Cloud OCR
- Open the Tencent Cloud OCR console and activate the text recognition service.
- Prepare your Tencent Cloud SecretId and SecretKey.
- Select Tencent Cloud OCR in WiseMindAI and enter the credentials.
- Enter a Region if needed, such as
ap-guangzhou; otherwise, leave it blank. - Save and test the recognition result.
Test OCR directly
After configuring an online OCR model, select Test. WiseMindAI recognizes text from a built-in image and displays the text and line count, without requiring you to import a production document.
If authentication fails, check the key. If WiseMindAI cannot connect, check the network and endpoint. If the result is empty, try another model or retry later.
Recommendations
- macOS users can start with Apple OCR because it requires no configuration and has no online service fee.
- Scanned documents, complex layouts, and tables require stronger recognition. Compare the results from Baidu Qianfan and Tencent Cloud.
- OCR can still produce errors. Check names, numbers, and specialized terminology against the source image.
- Free quotas, supported image sizes, and pricing for online OCR may change. Refer to the provider console for current details.