Recap is an open-source browser extension that lets you easily summarize any portion of any webpage with ChatGPT. It works by using ChatGPT, a large language model from OpenAI, to generate a summary of the selected text. This can be helpful for quickly getting the gist of a long article or for saving time when you need to research a topic.

To use Recap, simply install the extension and then highlight the text you want to summarize. Click the Recap button and the extension will generate a summary for you. You can then customize the summary by adding or removing text, changing the style, or adjusting the length.

Recap is a valuable tool for anyone who needs to quickly summarize text or who wants to save time when they need to research a topic. It is easy to use and produces high-quality results.

 

Here are some examples of how Recap can be used:

  • A student can use Recap to summarize the key points of a long article for a research paper.
  • A journalist can use Recap to quickly get the gist of a news story.
  • A businessperson can use Recap to save time when they need to research a topic for a presentation or report.
  • A language learner can use Recap to improve their reading comprehension.
  • Anyone can use Recap to quickly and easily summarize any portion of any webpage.

 

Overall, Recap is a powerful and easy-to-use tool that can help you to quickly summarize text or save time when you need to research a topic. It is a valuable tool for students, journalists, businesspeople, language learners, and anyone else who needs to read and understand text.

Chris: Hi there, please introduce yourself.

Yaozhi: Hello, my name is Yaozhi Wang, an indie hacker and Founder of Recap.

The concept behind Recap originated on a day when I stumbled upon an intriguingly lengthy article that I didn’t have the time to read in its entirety. Days later, when I returned to finish it, I realized I couldn’t recall the details of what I had read previously, although I remembered where I had stopped. This sparked the idea for a tool that could summarize content from the beginning up to a certain point – a feature that inspired the name “Recap”.

Chris: Recap is a Chrome extension designed to summarize web content. Could you explain in detail how the extension uses ChatGPT to perform these summaries, and what makes this approach effective in summarizing web content?

Yaozhi: Recap provides two modes to use ChatGPT:

1. ChatGPT webapp: For this mode, users are required to sign into the ChatGPT and keep the relevant tab open while using the service. The major benefit of this mode is that it is totally free, it just uses the chatgpt on behalf of the user.

2. OpenAI API, this alternative requires users to obtain and input their personal OpenAI API keys. With this method, there is no need to login ChatGPT, but users will be charged by OpenAI for the API usage.

These approaches are ideally suited for a browser extension that provides a seamless experience as users read the web pages.

Chris: The ability to split long webpages into manageable chunks and then summarize them is a unique feature. How does Recap determine where to split the content, and can users customize this process to better suit their specific needs?

Yaozhi: For a well-organized article, there should be different levels of headings to facilitate good readability. Recap capitalizes on this by dividing the article in a natural manner according to HTML tags: h1 – h6. Moreover, Recap offers customizable options for splitting the article into various levels.

Firstly, there is an application-side configuration that controls whether Recap should parse all pages by default; this can be set to either ‘Active’ or ‘Passive’ modes. When ‘Active’ mode is set by the user, Recap is enabled by default; it is disabled when in ‘Passive’ mode.

Secondly, domain-level and page-level toggles are available. Users can enable or disable Recap for the domain or page they are currently viewing by clicking the extension icon in the Chrome menu bar. The color of the icon will also indicate the current setting, which conveniently allows users to manage the feature.

Lastly, users have the option to customize the actual split points. They can select the container tag (e.g., <article>, <body>), exclude certain tags (e.g., <header>, <nav>), and choose the specific heading tags (from <h1> to <h6>). These settings can also be applied at the application level, domain level, or page level. These settings require the user to have some knowledge about HTML, but our default settings should be suitable for most of the articles on the web.

Chris: Customizable summaries are a valuable feature. Can you elaborate on the level of customization users can achieve with Recap, including summary length, style, and the specific information they want to include in their summaries?

Yaozhi: Customize summary is implemented by allowing users to create their own prompts. Through the use of diverse prompts, users can specify their desired output from ChatGPT, such as the language and style. We provide default prompts for some of the most commonly used scenarios, and we also offer suggested high-level prompts like ‘CoD‘ (Chain of Density), which can generate much-improved summaries for lengthy texts.

Chris: Professionals can save time by summarizing reports and work documents with Recap. How can this tool help professionals in various industries, and what impact does it have on their work efficiency and productivity?

Yaozhi: Professionals often deal with a deluge of information and lengthy reports that can consume a significant portion of their workday. Recap can distill these documents into concise summaries, freeing up valuable time that can be redirected towards other tasks. Recap aids in better understanding complex materials, which is invaluable for preparing for meetings, presentations, or making informed decisions without needing to sift through detailed documents. And Also Recap can help a lot in cross-Industry efficiency: Whether it’s legal professionals condensing case files, healthcare workers summarizing medical research, or marketers analyzing consumer feedback, Recap improves efficiency by providing quick insights across various sectors.

Chris: What types of websites or online content are most commonly summarized by Recap users, and are there any particular content categories where this tool has proven to be exceptionally useful?

Yaozhi: Sorry, since Recap is an open-source project, we didn’t track any usage data from users.

Chris: With the ever-evolving landscape of web content, how does Recap plan to adapt and stay innovative to meet the changing needs of users seeking efficient summarization tools?

Yaozhi: We are staying aligned with OpenAI’s advancements by regularly updating its system with the latest features and improvements. We also keep an eye on the latest research in natural language processing and text summarization to implement cutting-edge techniques that can complement the fundamental OpenAI algorithms.

We also plan to extend Recap from Chrome to other web browsers, so that it can serve more users. Additionally, we may support other AI models such as Bard, Claude.

Chris: As Recap evolves, are there any upcoming features, improvements, or developments in the pipeline that users can look forward to, and how will these enhancements enhance the summarization experience?

Yaozhi: The auto-split feature is the most standout aspect of Recap. We will continue to focus on helping users parse web content to summarize it more efficiently.

The algorithm that splits content based on h1 – h6 tags is most suitable for blog posts or articles. However, it may not work as well for other types of websites. We plan to develop site-specific adapters to make it useful across different types of websites, such as Reddit, GitHub, and Quora.

In another scenario, particularly on social media platforms like Twitter or Facebook, when a link is shared, a Link Preview is automatically generated to exhibit some metadata from the target URL. Recap can facilitate users by creating a summary directly from this preview, eliminating the need for them to navigate to the actual page.

You can see our roadmap on github.

Chris: Lastly, could you explain the business model behind Recap, including any pricing structure or subscription plans, and how this aligns with the value it offers to users looking to streamline their online reading and summarization?

Yaozhi: Recap is open-source and free to use. Currently, we do not have a business model. In the future, we may consider offering a lifetime deal for the most advanced features as part of a premium plan, but the code will remain open-source.

Chris: Thanks for being with me, any last words? Where can our readers follow you?

Yaozhi: As mentioned before, Recap is free, so I don’t have any budget to promote it. I really appreciate Generative Courier for conducting this interview.

For those keen on staying updated with Recap, be sure to follow for the latest product news, or connect with me at my personal Twitter.