Famously the Coin team of AI model researchers and engineers at Chinese e-commerce giant Alibaba has introduced a major expansion to its CoinDeep research tool, available as an optional feature a user can activate on web-based CoinChat (ChatGPT’s competitor).
This update allows users to create not only comprehensive research reports with systematic references, but also interactive web pages and multi-speaker podcasts—all within 1-2 clicks.
This functionality is part of a Proprietary releasediffers from many of Kevin’s previous open source model offerings.
While the feature relies on open source models Qwen3-Coderfor , for , for , . Qwen-imageand Qwen3-tts To power its core capabilities, end-to-end experience—including research execution, web deployment, and audio generation—is Hosted and operated by Kevin.
This means users benefit from streamlined, integrated workflows without the need to build infrastructure. That said, developers with access to the open-source model could theoretically replicate similar functionality on private or commercial systems.
This update was announced by a team official x account (@baba_qwain) Today, October 21, 2025, stating:
“Queen Deep Research just got a major upgrade. Now it not only generates reports, but also a live webpage and a podcast—powered by Queen3-Coder, QueenImage, and Queen3-TTS. Your insights, now visual and audible.”
Multi-format research output
The basic workflow begins with a user request within the QuenChat interface. From there, Kevin collaborates by asking clarifying questions to shape the research scope, pulls data from web and government sources, and analyzes or resolves any inconsistencies he finds — even developing custom code when needed.
a Demo video posted by Kevin on X Kevin goes through this process on Chat using American SaaS Market as an example.
In it, QWen pulls data from multiple industry sources, points out discrepancies in market size estimates (eg, $206 billion vs. $253 billion), and highlights ambiguities in the U.S. share of global data. The assistant comments on the differences in scope between the sources and calculates a 19.8 percent compound annual growth rate (CAGR) from 2020 to 2023, providing contextual analysis to back up the raw numbers.
Once the research is complete, users can click on it "Eyeball" icon below Output Results (see screenshot), which will bring up a PDF-style report in the right-hand pane.
Then, when viewing the report in the right-hand pane, the user can click on it "create" button in the upper right-hand corner and choose from the following two options:
"The web giant" which produces a Live, professional-grade web pageautomatically deployed and Hosted by Kevinusing Quen3-Coder for texture and QuenImage for visuals.
"The podcastfor , for , for , ." Which, as it describes, produces audio The podcastalso featuring dynamic, multi-speaker narration produced by Kevin 3-TTS Hosted by Kevin For easy sharing and playback.
This enables users to instantly convert a single research project into multiple formats – written, visual and audible – with minimal additional input.
The website includes inline graphics produced by Kevin Image, making it suitable for use in public presentations, classrooms, or publishing.
The podcast feature allows users to select 17 different speaker names as hosts and 7 as co-hosts, although I couldn’t find a way to preview the voice outputs before selecting them. It’s designed for deep listening on the go.
There was no way to change the language output that I could see, so mine came out in English, as did my reports and initial prompts, even though Kevin LLMs are multimodal. The voices were a bit more robotic than other AI tools I’ve used.
Here is an example of a web page I created On the commonalities of authoritarian regimes throughout historyfor , for , for , . Another one on UFO or UAP sightingsand below this paragraph, a podcast on UFO or UAP sightings.
Although the website is hosted via a public link, the podcast must be downloaded by the user and cannot be publicly linked to, from what I can tell in my brief use so far.
Note the podcast is very different from the original report—not a straight-through read-through audio version, but rather, a new format in which two hosts discuss the subject and use the report as a jumping-off point.
The webpage version of the report also does not include the new graphics in the PDF report.
Comparison with Google’s Notebook LM
While many early adopters have welcomed the new capabilities, comparisons to other research assistants have surfaced — notably Google’s. Notebook LMwho recently came out of Beta.
AI commentator and newsletter writer Motte (@kimmonism) noted on X:
“I’m really grateful that Kevin provides regular updates. It’s great.
But trying to create a Notebook LM clone inside the QWEN-3-MAX doesn’t look very promising compared to Google’s version.
Although Notebook LM is built around organizing and querying existing documents and web pages, Kevin Deep is more focused on research. Generating new research material from scratchcollecting sources from the open web, and presenting it in multiple ways.
The comparison shows that while the two tools overlap in general concept—AI-AISISTED research—they target approaches and user experience.
Availability
Kevin Deep Research is now live and available via Kevin Chat App. Access can be achieved with the feature The following URL.
As of this writing no pricing details have been provided for QWEN3-MAX or specific deep research capabilities.
What’s next for Kevin Deeper?
By combining research guidance, data analysis, and multi-format content creation into a single tool, QuenDeep Research aims to streamline the path from idea to published output.
The integration of code, visuals, and audio makes it particularly attractive to content creators, educators, and independent analysts who want to scale their research in web- or podcast-friendly formats without changing platforms.
Still, comparisons to more specialized offerings like the Notebook LM raise questions about how the Kevin’s overall approach stacks up on depth, precision, and refinement. Whether the power of its multi-format implementation outweighs these concerns is a matter of user preference—and whether they value single-click publishing over tight integration with existing notes and content.
For now, Kevin is pointing out that research doesn’t end with a document — it starts with one.
Let me know if you want it revamped or tailored for a particular audience.