PDF / Image Text Extractor
PDF files usually come in two formats:
- Single-layer PDF: each page is just an image. Every character you see is actually pixels, so you cannot select or copy it, nor can you search for it.
- Double-layer PDF: the underlying layer is an image, with a transparent recognized text layer overlaid on top. It looks the same as a single-layer file, but you can select text, copy and paste, and search with Ctrl+F. It is generally generated by OCR software.
For double-layer PDF files, the system can directly read the text content for use in AI writing.
For single-layer PDF files, you need to first use this tool to recognize and extract the text from the images and save it as reference material. This process is commonly called Optical Character Recognition (OCR).
This tool uses technical interfaces from third-party providers such as Alibaba Cloud to extract text from single-layer PDFs.
Note:
- The completeness and accuracy of OCR-extracted text depend on various factors such as font clarity and layout complexity in the original file. Even the most advanced OCR technology currently cannot perfectly extract all content completely and accurately. Therefore, depending on how rigorously you require the reference materials used for AI creation, you may need to further verify the extracted content.
- OCR technology providers usually charge by the number of pages, with each page potentially costing a few cents, making it a relatively expensive service. For trial customers, this system does not provide a free quota.
- You can also use many other OCR tools or products to extract text from single-layer PDF files.
Article Polisher
Usually, articles created by AI always more or less bear characteristics of machine creation, and differ from human-written works in aspects such as grammatical style and structural format.
During the AI creation process, this system has already used various technologies to optimize and eliminate these machine-creation characteristics as much as possible. However, if you are still not satisfied with the content of AI-created books, you can use this tool for further processing and polishing.
In addition to polishing the content of book chapters, you can also paste other text into this tool for polishing.
Note:
- You can directly import a chapter of a book created by AI into this tool, or directly paste other text, but the total length of text polished each time is limited to 10000 characters;
- The polished content can be exported as a Word or Markdown file, but it cannot be written back directly to the book to replace the original chapter content (this may damage the original book's style and structure);
- The polishing tool also consumes a certain amount of Token quota (you can view usage records in the AI call logs).
- This tool mainly uses the Humanizer technology developed by Siqi Chen.