Limitations/Project Future

1. Lack of Historical Pricing The market page currently provides only spot pricing, limiting users to real-time token values without access to past performance or trends. This restricts users' ability to analyze historical data for better decision-making. The market price movement indicators currently do not function, and are kept as a placeholder.

   Improvement Opportunity: We plan to integrate historical pricing data into the market page to enhance transparency and provide users with a more comprehensive understanding of token trends. This feature will be implemented upon receiving the API token from Pragma Oracle.

2. AI Agent Hallucination The AI agent on the platform is capable of answering user questions about tokens; however, due to the lack of a paid API token that allows it to search the internet, the agent may hallucinate or fabricate information. This could lead to inaccuracies in the data provided to users.

   Improvement Opportunity: Acquiring a paid API token will enable the AI agent to access real-time internet data, significantly reducing the risk of hallucinations. This enhancement will improve the reliability and accuracy of the AI agent's responses, making it a more dependable tool for users.

3. Limited Bot Deployment Currently, the Teleswap bot has only been deployed on the Telegram test server, which limits its accessibility and real-world usage. This restricted deployment prevents users from fully interacting with the bot in live environments, thereby impacting its practical utility.

   Improvement Opportunity: Deploying the bot to the main Telegram server is a priority for future improvements. This will ensure full functionality, broader accessibility, and the ability to cater to a wider user base in real-world scenarios.

4. Limited Token Support and liquidity on Sepolia Since this is a proof of concept running on Sepolia, we currently support swaps only with USDC, ETH, and STRK. This limitation is due to the restricted token liquidity available on the Sepolia testnet.

   Improvement Opportunity: A mainnet deployment would address this issue by supporting a wider range of tokens, leveraging the significantly greater liquidity available on the Starknet Mainnet. This would enhance the platform’s utility and provide users with more trading options.

5. Token List Approval Method The token list approval method needs to be more scalable and flexible. Currently, the process involves hardcoding token lists into the system, which limits the ability to easily add or remove tokens and complicates maintenance.

   Improvement Opportunity: To enhance scalability and flexibility, the token list should be fetched dynamically via an API. This would allow the system to call an external service that provides up-to-date information about approved tokens, removing the need for manual updates and improving overall efficiency. By integrating an API, the system can automatically handle token additions or removals, making it easier to scale and manage token approvals without requiring code changes.

6. Argent SDK Limitations The Argent SDK is currently in its alpha stage. These challenges have hampered our development cycle, leading to delays in implementing certain features and functionality. As the SDK evolves, we anticipate that these issues will be resolved, allowing for smoother integration and development moving forward.

7. Lack of Balance Awareness in AI Currently, the AI agent does not have the capability to access or interpret users' wallet balances. As a result, user commands such as "swap all my ETH for STRK" or "sell half my USDC" cannot be executed, limiting the AI's ability to provide an intuitive and seamless user experience for token swaps.

   Improvement Opportunity: To overcome this limitation, we propose integrating a balance query tool and implementing tool chaining within the AI system. This enhancement will enable the AI to dynamically fetch wallet balance data and process commands with precision. For instance, users will be able to issue commands like "swap all my ETH" or "sell half my USDC," and the AI will accurately interpret and execute these actions. This upgrade will significantly improve user experience, making token management more intuitive and the platform more versatile.

8. Future Voice Input Feature The code is already implemented and needs to be debugged. It will make the platform more accessible and hands-free, users can interact with the AI via voice commands. This feature enhances usability, especially for mobile users.

Conclusion

We welcome any contributions to the project, particularly in any of these aspects. Feel free to fork the project or make a PR.

Thank you for reading!