https://x.com/kevin0connell/status/1985392831641624893
I would really appreciate RTs on this - especially if you find this cool 😀
tl;dr, this is a library that will make an MCP server for any website you find on the web 🫡
Its very easy to be discouraged at times over the ups and downs of everyday life
with that being said - every day I wake up I am incredibly blessed to be surrounded by a group of people who love, cherish, and support me at every turn
will never take that for granted, and neither should you 😀
Gonna be in SF from Saturday night to Wed morning
If you’re building in AI (or hiring), would love to chat 🙂
A few things i have built:
- Helped scale @neynar
- Built a search engine for the Farcaster network
- Created a customer support agent called Agentscan that uses LLMs, RAG, and web scraping (video demo here)
https://x.com/jpoGMI/status/1864419482971849081
Most of my work is in Typescript and Python as a full stack engineer.
HMU 🙂
🚨 Developers who have used the OpenAI Agents SDK (Python preferred, but also will take Typescript 🙂) 🚨
I am starting to spend time working on a library that helps you better locally test/build a agent - would love feedback on this 🙂
high level:
1.) within a local UI you should easily be able to load the agents you have in your repository, play around with the agents instructions/temperature, chosen model, etc, and test responses.
2.) you should easily be able to test the agents you have using pytest - same concept as 1. ideally this should work within a github ci also.
3.) stretch goal (need to think more about this more), you can easily scaffold a tool that you can use with a agent (think a sdk, api, etc).
These are some things I personally would find useful, as I haven't found a non vendor locked in option for the above. If you think this would be useful too, or you want similar things that aren't the above, please let me know 🙂
