On June 6, 2025, Mercari co-hosted a meetup event for users of the AI code editor Cursor at Mercari’s Roppongi office in Tokyo. Cursor Meetup events are held all across the world. For the Japanese version of this event, Mercari organized a jam-packed program including lightning talks about use cases of Cursor at different companies and a special message for Japanese users from the Cursor team.
Mercari employees co-hosted this hybrid event, which attracted over 6,000 participants in-person and online and highlighted the strong interest in AI-powered development tools across Japan.
See below for our report on the highlights of each session and Mercari’s session.
Overview of event and session highlights
First, let’s take a look at the highlights of each of the five sessions in the program.
Cursor Agent Tips for Non-Engineers (by Miyacchi, EXPLAZA, Inc.)
“I will never manually create a slide presentation ever again,” stated presenter Miyacchi.
Miyacchi talked about a novel approach to Cursor by using it as an AI-powered notepad, which can assist those with no engineering experience.
He then presented the audience with the results of a workshop in which participants used Cursor to help them write fanfiction about their favorite celebrity or character. All 350 participants said that they were able to get the hang of Cursor despite having no technical background.
“Value is shifting from being defined by skill and is becoming more about having the will to try”
This session revealed the possibilities that can be unlocked by the power of AI.
Cursor x Genomics: Between Evolution and Restriction (by Prof. Masaru Koido, The University of Tokyo)
In the field of research, leveraging Cursor has very clear advantages and pitfalls. Professor Masaru Koido of The University of Tokyo shared his frank opinions on the pros and cons of Cursor in this field.
He stated that, “Cursor is undoubtedly accelerating machine learning and research that does not involve personal information or other sensitive information.”
While he praised Cursor for this potential advantage, he also touched upon the realistic limitations of the technology, namely, that Cursor cannot be used for genome data analysis, which does involve personal information. Using the latest policy published by the National Institutes of Health (NIH) as an example, he explained the strict limitations and harsh reality of AI in research work.
When handling extremely sensitive information such as genome data, simply inputting this data into an AI engine counts as providing information to a third party, which is strictly prohibited. This creates a dilemma for researchers who are unable to use the latest AI tools despite the benefits they could afford. Professor Koido’s talk highlighted the fundamental issue that many researchers face in striking a balance between the leveraging the potential of technology and ensuring privacy.
Vibe Coding 101: Cursor 15-Min Cooking (Naru)
Meet Naru, a go-getter who spends 200,000 JPY a month on AI tool subscriptions. Naru planned to demonstrate “vibe coding,” a coding method touted by Andrej Karpathy, but had to suspend the session due to connection issues.
However, Naru found a way to turn the situation around. He deftly dove into a casual talk about how technical problems like the one he was experiencing are actually quite interesting, and entertained the audience by pointing out that we are pretty much powerless now without the internet.
Naru taking this unexpected issue in his stride really embodied the unique nature of a live event.
Operational Rules and Insights When Using Cursor at a Security SaaS Company (Tetsu Takizawa, Cloudbase)
Tetsu Takizawa, who works for security systems company Cloudbase, shared some practical advice about using Cursor safely at a company.
He started with a warning about the dangers of MCP servers (linking Cursor with external services). Using an integration tool made by a developer with ill intentions could result in confidential information being stolen (tool poisoning attacks), or new malicious features being added to a secure system after release (rug pulls).
He stated that Cloudbase has one simple rule—never input user information into an AI engine. However, sticking solely to that rule for all development would negatively affect efficiency, so Cloudbase implements practical measures such as only using reliable, official tools and fixing versions to prevent unauthorized updates.
“It’s important to leverage AI not only to pursue convenience but also in consideration of company rules, laws, and, above all, the end user.”
Cloudbase’s balanced approach of enhancing productivity while ensuring security was impressive.
Overcoming the Barriers to AI Tool Implementation in Major Corporations: CyberAgent’s Cursor Utilization Strategy (Günther Brunner)
Günther Brunner of CyberAgent talked about the real-world issues that can occur when starting to use AI tools at large corporations, and offered some possible solutions.
CyberAgent is a major organization with over 10,000 employees, 3,000 of which are engineers. It also comprises over 100 subsidiaries. While their free-spirited corporate culture has been highly innovative, it has posed significant barriers to company-wide tool implementation. The same has applied to promoting Cursor at the company.
To solve this problem, Brunner has implemented a clever strategy leveraging “external influence.” In February of this year, CyberAgent hosted the external event AI Code Agents Festival. This event garnered a huge response with over 20,000 views online, prompting a flood of direct messages from previously indifferent CyberAgent employees asking if they could use Cursor as well. It was a strategy that cleverly leveraged the psychology of people being drawn to something when it gains external attention.
メルカリ2,000人展開の実践事例 導入から活用へのリアルな道のり
Practical case study of Mercari’s deployment to 2,000 employees: The real journey from implementation to utilization
Click here to view the presentation slides. (Available in Japanese only)
Cursor usage at Mercari
The first presenter from Mercari was @Kuu, who spoke about the current state of usage of Cursor and other AI tools at Mercari. @Kuu started by introducing Mercari’s approval process. Many attendees were shocked at how simple the process is.

“Employees can just send ‘I want Cursor’ to the IT Service Agent, one of our internal Slackbots, and receive a simple ‘OK’ as a response.” Naturally, no application forms or any reason for the application are required. After applying, Cursor is also registered to Okta, meaning that registration, including SSO, is automated as well.
Having such flexibility has given rise to an active Cursor community at Mercari of over 500 members. Knowledge is shared daily in the Cursor “waiwai channel” (a casual chat channel). In addition, engineers are provided with a wide range of other options to choose from, including tools like GitHub Copilot, Devin, and our in-house LiteLLM tool.

However, in spite of this seemingly hitch-free implementation phase, there were also issues.
“You can only unleash Cursor’s full potential when it has access to information from GitHub and Jira.”

As @Kuu pointed out, addressing security risks is an important topic when introducing tools. To address security needs, Mercari adopted a twofold approach: an official MCP server verified by the Security Team, and a set of in-house MCP servers tailored to internal needs.

These servers are consolidated in a single repository to create a system that promotes bottom-up development while ensuring security.
Becoming an AI-Native organization
Next, @komatsu introduced one of Mercari’s bold initiatives.
“You are forbidden from writing any code yourself for a week.”
This was the only rule of Mercari’s PCP LLM Week, in which 50 engineers took part.

This rule shocked the audience when they heard it. Even typos and one-line changes had to be handled by Cursor—wasn’t that too strict? While this rule might seem restrictive at first glance, there was a reason behind the madness.

“By setting a specific timeframe and turning the challenge into an event, the organization was able to make a bold decision to ‘accept a short-term decline in productivity for the sake of mid- to long-term benefits.’”
And, the results were even better than expected. Even more impressive than the scores of 92% satisfaction and 96% intention to continue using Cursor was the change in participants’ mindset.
“I had never used Cursor before participating, but now I can’t live without it!”
“Now, I think about how I can use Cursor for each of my tasks.”
The feedback from participants spoke of a fundamental transformation that went beyond merely implementing the tool.
Next, @komatsu, who planned and managed PCP LLM Week as engineering manager, shared three important lessons from the experience.

First was the importance of enforcement. Even in an organization like Mercari where many engineers are highly proactive, designating a period in which the organization is encouraged to try out AI created an opportunity for all employees to seriously engage with AI.
The next lesson was encouraging information sharing. Actively sharing knowledge via Slack has produced results beyond expectations and created a community for continuous learning.
Lastly, @komatsu mentioned the standardization of skill levels. When all employees share the same experiences, the AI literacy of the whole organization improves, and this creates a foundation for employees to discuss topics using a common language.
“At Mercari, we use a variety of AI tools, including Cursor, as a normal part of our everyday tasks. We want to implement more of these initiatives and continue our efforts as an organization to become AI-Native with an ‘All for One’ approach.”

Japan records the third highest number of Cursor users around the world
To close out the event, Cursor Ambassador Kinopee took to the stage. Known as “Mr. Kinopee” among overseas AI professionals, he is a central figure in promoting Cursor in Japan, having authored books on Cursor and contributed to feature articles in magazines.
For this event, Mr. Kinopee played a video message sent by the Cursor team. The messages from Ryuu, the design lead, and Michael Toll, Cursor CEO, conveyed gratitude and anticipation to Japanese users. Then, the revelation of a certain fact caused a stir in the venue.
“Japan ranks third in the world for the number of Cursor users.”
When Mr. Kinopee attended a Meetup in San Francisco last year, Japan was reportedly ranked fourth in the world. “The ranking went US, China, India, and Japan. To be honest, I was a bit disappointed that Japan wasn’t higher,” said Mr. Kinopee. Finding out that Japan had climbed to third in the world made me again realize the passion of the Cursor community in Japan.
Conclusion
Cursor Meetup Tokyo was an event filled with enthusiasm from start to end. Over 6,000 people participated, both in person and online. This large number indicates that Japanese engineers have high expectations for AI tools and feel positively about actively adopting new technology.
We hope that sharing Mercari’s efforts to move beyond simply implementing AI tools to effectively utilizing them will provide some hints for many companies facing similar challenges.
See you at the next event!
Original Article: Nana Naka / English Translation: Mercari Global Operations Team
Related job positions
Here are some of our open positions!
-
Software Engineer, Web – US App
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
AI/Data Product Engineer – Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Software Engineer, Cloud Networking – Mercari (Japanese Speaker)
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Software Engineer, Backend – Mercari Mobile (モバイル領域新規事業)
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Corporate Engineer, IT Service Management- Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
IT Infrastructure Engineer – Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Technical Project Manager (Corporate DX) – Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Systems Engineer (Corporate DX) – Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Software Engineer, Cloud Networking – Mercari
Office: 東京・六本木オフィス
Company/Business:
-
Product Engineer,Backend – Mercoin
Office: 東京・六本木オフィス
Company/Business: メルコイン
Direct you to a careers site
Related job positions
Here are some of our open positions!
-
Software Engineer, Web – US App
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
AI/Data Product Engineer – Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Software Engineer, Cloud Networking – Mercari (Japanese Speaker)
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Software Engineer, Backend – Mercari Mobile (モバイル領域新規事業)
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Corporate Engineer, IT Service Management- Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
IT Infrastructure Engineer – Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Technical Project Manager (Corporate DX) – Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Systems Engineer (Corporate DX) – Mercari
Office: 東京・六本木オフィス
Company/Business: メルカリ
-
Software Engineer, Cloud Networking – Mercari
Office: 東京・六本木オフィス
Company/Business:
-
Product Engineer,Backend – Mercoin
Office: 東京・六本木オフィス
Company/Business: メルコイン
Direct you to a careers site