Farcaster is a playground for crypto experiments—I find myself discovering and trying something new every week, and more often than not, I’m surprised. This is my attempt to log my experiences.

Earlier this year, I started coding again after a long time, of course, with a lot of help from Cursor. One idea that’s fascinated me about new AI bots is having an intelligent companion that helps you with your tasks. Now, there is one task that is ubiquitous in crypto. Almost everyone in crypto trades or invests in tokens.

Ask Gina is a wallet you can talk to. You can ask her for prices, trends, what tokens are hot right now, and you can ask her to do transfers and swaps for you. And it will do all of that for you and more.

To get started, set up a Privy wallet on AskGina.ai, fund it with USDC on Base, and you’re ready to start your prompting journey.

Don't be me though. I had a false start, primarily because I ignored the instructions and sent across ETH 🙈 After that I tried asking gina to transfer some ETH to a different wallet which she failed to. I'm guessing presently Gina can operate with only ERC20 tokens. Anyways, it does have a export private key feature so am assuming you can import the wallet into your regular wallet and still use it. The onboarding was otherwise fairly seamless for me.

Now, there’s no doubt—an agent that can transact on-chain on your behalf is impressive. The experience however that came close to magic for me was a research prompt.

what is the potential price impact we can see on CLANKER TOKEN once it gets listed on coinbase? support insight with data of past listing in last 12 months

I'm skipping the full response in interest of keeping this post short. The response was significantly better than what I got from Perplexity Deep Research (which was riddled with inaccuracies) and comparable with that of Chat GPT's Deep Research. A few other prompts that I tried were around researching Aave's revenue and pricing trend and identifying the biggest risks in my portfolio. The responses were generally helpful. This is one use case I am excited about. The ability to do a deep research on a token or category, brainstorm on different trading strategies, and then put them on automation.

There are so many different ways the product can evolve. An onchain research assistant that can trade for you . A trading assistant that can help you with your research. The difference is subtle but significant.

framedl is one my favorite app on farcaster. A wordle game a day right in a frame. I have played over 300 games now. But this week for the first time I tried out the framedl arena.

As the name suggests, you have an arena with a fixed number of spots where people can join. There are 5 games. And people are ranked based on their average score. While the games are played asynchronously, you can track real time update of the score which makes it super thrilling.

One of the reasons why I couldn't post Farcaster Fridays last two weeks. I was knee deep picking up React and trying to learn to code a frame. Introducing Reconnect.

Reconnect helps surface old frens you made on Farcaster but lost touch with. I’ve always found it a bit sad that algorithms optimize for engagement, not for the people you meet along the way.

Built this with a lot of help from Cursor, using React & Node. The data runs on Dune, so yeah... it’s slow (free-tier struggles). Hope you give it a try and reconnect with someone!

Rch on Warpcast

last month, started writing code again after a long time. my first frame. we made frens along the way - some we lost touch with. maybe it's time to reconnect. https://reconnect-frame.replit.app

Farcaster