Dev Diary, en

Kryll³ Dev Diary - 10th Edition: Early Access Opening

Welcome to the tenth edition of the Dev Diary, where we’re excited to share Kryll³’s remarkable progress, marked by the launch of our first early-access version!


Since November 5, around fifty users have been testing the platform’s new features in real-world conditions: SmartFolio, News Hub, Agent K, and the Reward Program. This early-access phase represents a key milestone in Kryll³’s development, enabling us to validate the stability and usability of these features through real user feedback.

  • Current Context of Kryll³:
    Kryll³ is a Web3 platform offering users powerful portfolio tracking, market analysis, and decision support tools, all enhanced by artificial intelligence. This initial version allows us to assess our hybrid infrastructure and test the performance of our KortexFlow AI model under real-world conditions.

Key Objectives of This Early-Access Phase

The primary aims of this early-access testing phase are to:

  • Validate the stability of our hybrid infrastructure (serverless architecture, dedicated servers, and integration with decentralized web3 services).
  • Test the capabilities of our AI (KortexFlow and version 1 of Agent K) and its performance in real scenarios.
  • Gather feedback from our first users regarding usability, bugs, and the utility of the features.

Highlights of Kryll³’s Platform Advancements

SmartFolio :

Screenshot of SmartFolio from the Kryll³ platform
  • Available Features:
    SmartFolio supports the Ethereum and Solana blockchains, as well as Layer 2 solutions like Polygon, Base, and Arbitrum, along with major CEXs: Bybit, Coinbase, Gate.io, and Kucoin in this very first version. Users can monitor their portfolios in a unified way, with aggregated views, detailed breakdowns by chain, and advanced filtering options. Additionally, they can interact directly with Agent K to obtain analyses of their portfolio based on personal data.

News Hub :

Screenshot of NewsHub from the Kryll³ platform
  • Content and Analysis:
    The News Hub aggregates news from French-speaking sources (Journal Du Coin, Cryptoast, CoinAcademy, Bitcoin.fr) and English-speaking ones (Beincrypto, Cointelegraph, Coindesk, TheBlock), providing comprehensive coverage of crypto news. Articles are analyzed in real time by AI models that assess the impact and sentiment (positive or negative) of each piece of news. Linked with Agent K, the News Hub allows users to request contextualized summaries (“summarize crypto news in light of the US elections”), extract specific information (“what’s the latest on CZ”), or get an overall analysis (“Is today’s news looking positive?”).

Agent K :

Screenshot of Agent K, the AI conversational agent on the Kryll³ platform

AI Capabilities:
In this initial version, Agent K is designed to offer a wide range of services based on real-time data. Equipped with multiple analytical tools and access to market and user portfolio information, it can respond to a variety of queries and provide detailed analyses to assist users in making investment decisions. Here are its main capabilities, with concrete examples:

  • User Portfolio Consultation:
    Agent K has access to the user’s full portfolio, including assets stored across different blockchains (Ethereum, Solana, etc.) as well as on centralized platforms (Bybit, Coinbase, etc.). This enables it to provide personalized insights based on the user’s holdings.

Example: “What are my top-performing assets over the past week?” or “Can you analyze my investments by comparing their performance on Solana vs Ethereum?”

Screenshot of the question "What are my best performing assets over the past week?" asked to Agent K, the AI conversational agent on the Kryll³ platform

  • Real-Time Market Analysis:
    Agent K can retrieve up-to-date information on prices, market cap, trading volume, and price fluctuations of various cryptocurrencies. This allows users to quickly get a snapshot of market conditions and spot opportunities.

Example: “Which tokens have seen the largest price increase over the last 24 hours?” or “What is the current trading volume of Bitcoin and Ethereum?”

Screenshot of the question "Which tokens have seen the biggest price rises in the last 24 hours?" asked to Agent K, the AI conversational agent on the Kryll³ platform

  • Crypto Project Research:
    Agent K can also provide in-depth information on specific projects, including their name, website, whitepaper, category, and social media activity. This helps users discover promising projects and better understand the cryptos they might want to invest in.

Example: “Can you give me more details about the Aave project?” or “What is Solana’s current level of social activity?”


  • Market Narrative and Category Analysis:
    Agent K can group and analyze tokens by thematic categories (such as “Real World Assets” or “Layer 1”), allowing users to identify trends and spot performing or underperforming categories.

Example: “Which tokens in the 'Real World Assets' category have a market cap under $40 million?” or “Who are the top gainers among DeFi tokens in the last 24 hours?”

Screenshot of the question "Which tokens in the 'Real World Assets' category have a market capitalization of less than $40 million?" asked to Agent K, the AI conversational agent on the Kryll³ platform

  • Opportunity Identification (Gem Findings):
    Agent K is capable of identifying “gems,” meaning cryptocurrencies that may be undervalued or showing signs of interest, based on personalized criteria such as volatility, market cap, and trading volume.

Example: “Find me low-value tokens on Bybit with a trading volume above $1 million” or “Which assets on Binance and Coinbase might be ‘hidden gems’ based on the past 30 days’ price trends?”


  • Crypto Rankings by Specific Criteria:
    Agent K can provide dynamic rankings based on criteria like market cap, gains and losses over different periods, or specific categories.

Example: “Show me the top 10 cryptocurrencies by market cap” or “Who are the biggest losers in Layer 1 tokens over the past 30 days?”

Screenshot of the question "Who have been the biggest losers in Layer 1 tokens over the last 30 days?" asked to Agent K, the AI conversational agent on the Kryll³ platform

  • News and Sentiment Analysis:
    Linked in real-time with the News Hub, Agent K can extract, analyze, and summarize the most important crypto news. Beyond providing simple summaries, it can assess the general sentiment of the news (positive or negative) and contextualize it according to relevant events or key personalities.

Example: “What is the general sentiment in the news today?” or “Give me a summary of crypto news related to the US elections.”


  • Contextual Interaction and Structured Reasoning:
    Agent K is designed to respond in a structured manner, synthesizing information and placing it in context. If a user asks for complex information, Agent K can detail and explain key points to facilitate decision-making.

Example: “Explain the potential impact of Layer 2 developments on my investments” or “Why is the crypto market down today based on available data?”

Screenshot of the question "Explain the potential impact of Layer 2 developments on my investments" asked to Agent K, the AI conversational agent on the Kryll³ platform

Since the early-access launch, hundreds of questions have been directed to Agent K, and feedback has been highly positive. Users particularly appreciate its ability to provide detailed explanations and make complex concepts more accessible. Responses are considered relevant, and real-time access to web3 data is also a strong point. However, a few reasoning bugs have been reported, particularly when Agent K “forgets” that it has access to real-time data. To address these issues, the team is actively working on an enhanced reasoning engine to increase Agent K’s stability and reasoning capabilities.

Screenshot of a node flow in KortexFlow, Kryll³’s in-house GenAI operating system

Reward Program and Gamification:

Screenshot of the Reward Program hub on the Kryll³ platform
  • Encouraging Engagement:
    Users can earn AI credits by sharing content on social networks through GALXE quests, which they can then claim on Polygon or directly on the DApp. By using the platform and holding KRL in their connected wallet, they accumulate XP and unlock levels, granting them access to unique procedural avatars and ranking them on a progression dashboard. Feedback has been very positive, even though a few bugs have been reported and fixed along the way.

Next Steps

  • Overall Stabilization: Based on community feedback, the team has initially focused on optimizing platform stability to ensure a smooth and reliable user experience.
  • Improvement of Agent K’s Reasoning Engine: To strengthen Agent K’s reasoning abilities and improve response relevance, we are also working on an enhanced reasoning engine. Version 2 is scheduled for next week and is expected to significantly boost Agent K’s stability and accuracy.

Conclusion

This early-access launch marks a critical milestone for Kryll³. With robust features like SmartFolio, News Hub, and Agent K, along with a motivating Reward Program, we are taking a major step toward a powerful and accessible crypto analysis and investment platform. The feedback from early users has been very encouraging, and their suggestions allow us to continuously refine our tools.

We would like to extend our heartfelt thanks to our testers, former Kryll (v2) clients, for their loyalty and involvement in this new chapter of our journey. Their engagement and feedback are essential in building a platform that meets the community’s needs. A special thank you also to our partner, Google, who supports us in constructing this extraordinary project that is Kryll³.

Author image

About Svein