From Twitter App to AI Tools: My Journey with covelline, LLC.

Hi there, I'm Ryohei, founder of covelline. I wanted to share a bit about our company as we celebrate our 12th anniversary. For those of you interested in what we do, here's a quick overview:

  • Founded by a bunch of tech enthusiasts from a technical college
  • Created an app that maintained the #1 spot in paid SNS apps on the App Store for 7 years
  • Developed cutting-edge projects involving AR/VR and object recognition for various industries
  • Currently filed a patent connecting data and AI as we prepare for the age of AI agents

The Origins of covelline, LLC.

covelline started with three of us who were classmates at Oyama National College of Technology. Back in college, we competed in a programming contest together and developed an AR-based navigation system in 2009, before smartphones were even widespread. While we only received a fighting spirit award (our demo booth was a bit confusing and our visuals weren't flashy enough), we were proud of creating what we felt was a truly innovative product.

A few years later, when we saw similar concepts appearing in luxury car navigation systems and smartphone maps, we became convinced of the potential of our technical skills and creativity. Though the three of us had gone our separate ways after college, we decided to reunite and form covelline.

The members of covelline

(We were invited to speak at our alma mater's 50th anniversary event, sharing the stage with Nobel Chemistry Prize winner Professor Hideki Shirakawa of Tsukuba University)

"feather for Twitter": Maintaining the #1 Position for 7 Years

Our flagship product was "feather for Twitter", a Twitter client app. Its meticulously refined features and UI/UX were highly acclaimed by users worldwide, including various media outlets. The app maintained the #1 position in the paid SNS app category on the App Store for about 7 years.

Top of the App Store

Unfortunately, in 2023, due to Twitter's policy changes, we had to discontinue the app along with other third-party clients. The news actually made "feather for Twitter" trend on social media, with many users expressing their disappointment. This experience motivated us to develop a successor, "feather for Mastodon."

Our Technical Projects Beyond Apps

Beyond our own apps, we've worked with various clients on interesting development projects. Our technical focus areas include:

  • Mobile app development
  • Architecture and development leadership
  • Modernizing existing codebases

We've also taken on projects using newer technologies like object recognition and AR/VR, such as:

  • An AR app using facial recognition for a confectionery manufacturer
  • VR cognitive training for a healthcare startup
  • Object recognition visualization with AR glasses for a major automobile manufacturer

We also collaborated with our alma mater, Oyama National College of Technology, on an AR app for educational settings, which led to us co-authoring a research paper.

Throughout these projects, we've combined our skills in software, hardware, and design to create effective solutions. After developing Twitter client apps in the platform's early days and working with AR/VR technology before it became mainstream, we're now focusing on products for the emerging AI agent era.

The AI Agent Era and the Crucial Technology It Demands

AI has already significantly impacted the software industry. One emerging trend is the evolution toward AI that plans its own actions and proactively uses various tools to achieve broader goals, rather than just performing single tasks.

Imagine an AI that, when given an objective like "improve our company's sales", analyzes the current business situation, devises improvement strategies, and implements them without requiring detailed step-by-step instructions—functioning almost like a human team member. While these AI agents are currently only practical in limited fields like programming, they're expected to be applied across various domains in the coming years.

In this future, what's crucial is that AI agents can access necessary information and, most importantly, that this information is in a format that's usable for AI.

When information is presented in an AI-friendly format, it's equivalent to making the world more visible to AI. Only then can AI precisely understand the real-world issues and human factors embedded in that information.

This is why we're currently developing and patenting technology related to "formatting information in ways that are usable for AI." We've named this technology "Eyes", as it gives AI better vision, and we're working on products that effectively solve real-world problems with it.

One key innovation is using AI itself to perform the task of "formatting information in ways that are usable for AI." This approach allows us to build high-speed, scalable systems without human work speed or cost becoming bottlenecks.

Specifically, our patent-pending system:

  • Uses AI to appropriately format human log data (behavioral information) collected from web and apps so AI can understand its essence
  • Passes these results to AI for effective classification and aggregation
  • Further processes this with AI to extract insights like problem areas and improvement suggestions

In this system, no human intervention is required from the input of human log data to the extraction of human-oriented insights. By effectively connecting humans and AI, we can build systems that operate without requiring human effort. We've named this technology "Eyes", as it gives AI better vision into human behavior and data.

đź‘€ Eyes

Currently, we're offering the first in the Eyes series: Eyes Outreach, an AI-powered service for automated measurement, analysis, and improvement of sales emails.

Eyes Outreach

For companies sending large volumes of sales emails or newsletters, it provides bulk sending capabilities plus:

  1. Measurement of email open rates and link click rates
  2. AI-powered automatic analysis and learning of which types of email content are effective for which recipients
  3. AI-generated email content optimized for each user segment

It maximizes impact by thoroughly streamlining email-related tasks while creating personalized subject lines and content. This powerfully assists sales teams operating with limited personnel.

We're planning to expand with AI-powered automatic analysis and improvement suggestion services for web and apps, as well as offering one-off development to add AI analysis functions to customers' products.

If you're interested, please feel free to contact us!