Inspiration
Staying active and relevant on social media has become more important than ever, especially with this year's bull run and the surge of online activity across commerce, politics, education, and more. What started as a simple desire for an easier, more efficient social media presence has now become a necessity for billions of internet users. With so many people depending on boosted posts and paid ads just to stay visible, the need for a smarter, more affordable way to manage engagement and growth is clearer than ever—and that's exactly what inspired this project.
What it does
Our AI agent takes full control of managing your Twitter presence—automating posts, handling DMs with customed toned responses from available template, running and managing Twitter Spaces, and following set targeted users, all without requiring any manual input. It’s a complete hands-free solution designed to keep your account active, engaged, and growing while you focus on everything else.
How we built it
The AI was built with typescript for the functionality, SQLITE for the data management, Docker was used to containerise the AI agent to access the agent locally on a separate computer, Typescript for the near contracts development, Ava for the contract local test, React for the frontend development and UI/UX, NodeJs and ExpressJs for the backend server creation to enable communication between the users inputs on the app to the awaiting parameters expected from the AI agent.
Challenges we ran into
- Bridging the understanding gap between remote team members slowed down collaboration.
- Integrating the AI agent with the NEAR-powered frontend proved more complex than expected.
- We ran into unfamiliar bugs, including communication issues between the frontend and smart contracts.
- Compatibility problems with Apple Intel Core (DARWIN x64) devices caused major roadblocks for some smart contract developers—requiring thorough debugging and unfortunately, device migration.
- Poor documentation for frontend development on NEAR created challenges, from unstable dependencies to frustrating wallet setup issues with the boilerplate.
- Ethereum wallet connection required appKit, which was missing from the NEAR plugin docs, adding another layer of difficulty.
Accomplishments that we're proud of
- We accomplished a full working project that works on near protocol with the UI seamlessly communicating with the backend.
- We accomplished to build an AI agent that works on the near protocol and being able to send and request data for both the AI and the contract
What we learned
- We learned how to use the near protocol wallet to communicate with smart contracts and sign transactions
- We can now call functions from the frontend on near protocol.
- we learnt the near smart contracts development terminologies
- we understand the technologies used by Hot Wallet and documentation.
What's next for NURA(Near Unified Realtime Agent)
- charge users per subscription for using the AI agent
- fully integrate the developed working smart contract
- develop the UI to accept more options and parameters to configure the AI agent
- develop the AI agent to be automated for Telegram and Discord.
Log in or sign up for Devpost to join the conversation.