LilysAI is reborn as Readray

Starting today, LilysAI is reborn as Readray.

For nearly three years, 1.3 million people around the world have used LilysAI to read and understand the material that matters in their lives: papers, lectures, videos, and reports. And many of them told us the same thing: LilysAI is not just a summarizer. It helps them understand the original in depth, without missing what makes it valuable.

To push that value as far as it can go, we gave it a new name and redesigned the product from the ground up. It is our biggest leap yet, built on everything we learned over three years of LilysAI.

Why Readray

Readray combines Read and Ray. Just as an X-ray reveals the structure hidden beneath the surface, Readray helps you see straight to the core of whatever you are reading.

Think of Iron Man’s HUD or the glasses in Kingsman. Readray, too, puts what you need right in front of your eyes, every moment you read, so anyone can read like a genius.

We ran a closed beta with hundreds of users across North America and Europe, and Readray proved itself with a remarkable 72% retention. Today, we are opening it to everyone.

What’s new

1. An AI lens that helps you, wherever you read

General-purpose AI was not built around reading. Switching windows, attaching files to a chat, and writing prompts all break your reading flow, so we are not using 100% of what AI can do in every moment we read.

Readray reads along with you right where you are, whatever you open in your browser or on your computer. It is a proactive agent built for reading that brings you what you need to understand, without being asked.

• Reading a paper: it turns the paper’s figures into interactive ones you can explore to understand them, and researches the line of prior work behind it.
• Watching a lecture: it captures the whiteboard, slides, and code on screen as screenshots, and explains the background knowledge you need to follow along.
• Reading health or other accuracy-critical material: it lays out numbers and deadlines in a table at a glance, and fact-checks whether the information can be trusted.
• Watching a podcast: it looks into each guest’s background and compares their main claims side by side.
• Reading about investing: it checks the current price of every stock mentioned and the speaker’s track record.
• Watching a product review: it puts each product’s specs in a table, captures every product shown in a video review, and checks other reviewers to tell you whether the review holds up.

Readray works in Chrome (Safari coming soon), and on Mac and Windows desktops.

2. Best-in-class source recognition

Give a general-purpose AI a video or a document and ask a question, and it often answers without actually reading the whole file, which makes it hard to trust. Readray has spent years building the technology to read sources properly.

• For videos, it looks at the scenes, not just the captions. Ask it to capture every product shown in a fashion review and gather them into one image, or ask about what was written on the whiteboard in a lecture, and it gets it right.
• For paper PDFs, it extracts every image and table without missing any.
• Every AI output is linked to the original, sentence by sentence. Click the source on any sentence to open that exact spot in the viewer, which puts concerns about AI hallucination to rest.

3. High-quality reports, tailored to each source

Readray recognizes what kind of source you are reading and, in one click, writes the report that fits it best. Unlike a typical summary, a Readray report keeps everything that matters in the original while taking the friction out of reading it.

There are two kinds of reports.

• Key points: for when you want a quick, easy read. Rich in visuals, and written so simply that a grade-schooler could follow.
• Deep research: captures everything important in the original. On top of that, its External Research section looks up the context you need to understand the source and lays it out for you.

Depending on the source, External Research looks into things like:

• The line of prior work behind a paper, and whether its results are significant
• The background knowledge and terms a lecture assumes, plus the world’s best lectures, textbooks, and materials on the same topic
• Fact-checks of claims and statements
• Exceptions and caveats the original leaves out
• How credible the speakers are
• Opposing arguments
• For tutorials, what has changed in the latest version and how to fix common errors

4. AI chat that knows everything you have read

Readray chat has grown into a modern agent with skills, memory, and model choice. On top of that, it knows every trusted source you have read.

• Much more is possible than before. Capture screenshots from a video, edit a report, build a slide deck: ask for whatever you need, grounded in your sources.
• Chat across everything you have read. From organizing folders to reviewing what you read recently, try tasks that span many sources.
• Use a wide range of skills. Type ‘/’ to bring up skills for study, research, investing, and more. Our former extended reports are now chat skills too.
• Memory remembers what you ask for again and again, and automatically folds it into the quality of your next reports.
• Pick any flagship model you like, including Claude, GPT, and Gemini.
• MCP is supported too. Connect your sources and context to other general-purpose AI agents.

5. Credit-based plans

Plans are moving to credits. This change is what lets you choose your own model. All of your existing usage is preserved, and nothing about the change puts you at a disadvantage.

6. Five themes, including dark mode

We added five themes, including the dark mode so many of you asked for. Pick whichever is easiest on your eyes.

7. A cleaner screen, so you can focus on reading

We simplified the screens so nothing gets in the way of reading. As part of that, many menus have moved. We have documented every change in the Help Center, so check there if you cannot find a feature you used to rely on. If anything feels inconvenient, let our support team (support@lilys.ai) know and we will improve it quickly.

Closing

Thank you for loving LilysAI. Researchers, students, investors, and business leaders: everyone who reads with us is a valued customer. We hope Readray helps you read and understand faster and more deeply, so you can achieve what you set out to do, sooner.

Readray is only just getting started. We will keep working until we build a world where anyone can understand anything in a flash, without friction or fear, like Iron Man’s HUD or the glasses in Kingsman. Please keep sharing your thoughts, and we will keep getting better.

The Readray team

Readray