title | layout | nav_order" |
---|---|---|
Getting started |
home |
1 |
AI Reader is a free, all-in-one desktop software to help researchers and students conduct research with ease in the digital age.
{: .note }
AI Reader is currently in the preview stage (see development).
Table of contents
{: .text-delta } 1. TOC {:toc}Latest release: {{ site.github.latest_release.tag_name }}: {{ site.github.latest_release.published_at | date: "%b %d, %Y" }}
{: .label .label-green }
<iframe width="560" height="315" src="https://www.youtube.com/embed/FQfvJ_zOxOE" title="AI Reader 0.1.0 Preview 90 Secs Tutorial" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>Or, download the video here
Latest version: {{ site.github.latest_release.tag_name }} ({{ site.github.latest_release.published_at | date: "%b %d, %Y" }})
{: .new }
Legacy MacOS supported
{{ site.github.latest_release.body }}
{% for asset in site.github.latest_release.assets %} {% if asset.name contains '.msi' %} [Windows]({{ asset.browser_download_url }}){: .btn}
For Windows 7 and above. Size: {{ asset.size | times: 1.0 | divided_by: 1024 | divided_by: 1024 | round: 2 | append: " MB" }} {% elsif asset.name contains '.dmg' %} [MacOS Universal]({{ asset.browser_download_url }}){: .btn}
For both Silicon and Intel Macs. Size: {{ asset.size | times: 1.0 | divided_by: 1024 | divided_by: 1024 | round: 2 | append: " MB" }} {% endif %} {% endfor %}
Visit [releases]({{ site.github.releases_url }}) for legacy versions.
AI Reader focuses on the research process, setting it apart from other tools for reference management, note-taking, or writing. This approach ensures it meets the unique needs of researchers.
AI Reader is an all-in-one solution that covers the entire research cycle, from collecting materials to processing, analyzing, and finally outputting your work. This makes it a versatile tool for researchers.
AI Reader offers a variety of tools, processors, and services with intuitive graphic interfaces. You can choose and use these out-of-the-box features to customize your research experience.
Your data is stored locally on your computer in SQLite database files, which can be opened by any third-party database applications. This ensures you're never locked into a specific platform.
AI Reader is completely free. There are no plans for an online version or a cloud service with subscription fees. The only costs you might encounter are for services like OpenAI, which require an API account. However, even these will be phased out as comparable free alternatives become available locally.
As a fellow researcher, I understand the challenges academics face when transitioning from traditional library-based research to a digital environment. AI Reader is designed to help you prepare, read, and analyze papers in a digital format, offering easy-to-use tools for this purpose.
At its core, AI Reader aims to eliminate repetitive manual tasks, allowing you to focus on the aspects of research that truly matter: exploring, thinking, analyzing, and expressing. Drawing from my years of experience building tools for my own research, I hope AI Reader continues to evolve and help others in similar situations.
The first released AI Reader module, AI Discussant, allows you to chat with AI about your research materials. It
currently utilizes OpenAI's GPT-3.5 Turbo model
and requires an
OpenAI API key to function.
- AI Discussant in AI Reader focuses on academic needs, delivering thorough, sophisticated, and detailed answers backed by textual evidence from your document.
- Compare AI Discussant with similar products and see the difference. Feel free to [report issues]({{ site.github.issues_url }}) if you find AI Reader providing an inferior answer about the same document.
- AI Discussant answers include page references to the PDF document, enabling you to quickly verify and/or closely read the referenced page.
- Click on the provided list of keywords to search through the PDF document effortlessly.
- ChatGPT embedding and chatting costs are estimated, so you only pay for what you use. There are no subscriptions, page limits, or middle-man services.
- After indexing your document, you can not only chat with ChatGPT about it but also use the document for other purposes (with upcoming features) such as semantic search. There's no need to index the document again.
- Export your chat history as a Markdown file for future reference.
Upcoming in v0.2.0 {: .label .label-yellow }
- An integrated module that simplifies the extraction, browsing, organization, and export of your reading notes in PDF files.
AI Reader is a one-man project, and after years of exploration, it has finally reached a stage where it can
realistically support all its planned features. With most of the infrastructure (database, backend, and frontend) in
place in v0.1.0 Preview
, new feature development will be quicker.
As a researcher rather than a full-time developer, I can't guarantee timely delivery of new modules. However, I aim for
a monthly release cycle for new modules until AI Reader reaches v1.0
. Individual features and bug fixes will be
addressed as quickly as possible in shorter update cycles.
A feature is a specific function in AI Reader, such as exporting to Markdown. A module is a collection of features serving a broader purpose, like AI Discussant, and takes longer to develop.
- v0.1.0 Preview - AI Discussant Module - May
- v0.2.0 Preview - Notes Organizer Module - June
- More planned modules and features to be announced.
- AI Reader is free but closed-source, developed in a monorepo, and cannot accept pull requests.
- The best way to support AI Reader is by [reporting issues and requesting features]({{ site.github.issues_url }}) that are useful to you as a researcher.
- Currently, there's no option to donate to AI Reader. In the future, I may consider accepting donations for a fixed amount needed to obtain a Microsoft EV code signing certificate for AI Reader. This would eliminate the security warning dialogue when running the unsigned release. However, it's optional for now. The Mac version is already notarized.