A solo-founder is a myth!
I used to think that to build a company, you had to have a lot of things figured out in advance, and if that wasn't the case, then something was off and you weren't ready ...
I'll be pitching @LunchFlow next Saturday Deen Developers's Demo day in London, come through if you're around to see what me and others on the batch have been cooking up, and my plan for world domination.
Built a remote MCP for @FeaturebaseHQ that I can share with my team at @LunchFlow, featuring:
- An OAuth Flow
- Different API Keys per user
- Control which tools each user get access to.
- Single command deployment to Cloudflare.
Plugins in Cursor are simply packaging other primitives like Skills, Subagents and MCP servers, and making it slightly easier to share them.
I look forward to an open standard that makes this tool independent, for those of us that don't want to be locked into Cursor.
This all started from a comment on a Github issue where people are looking for an alternative to @GoCardless 's bank account data, now that they don't offer their free tier to individuals any more. Helps validate the need for @LunchFlow !
https://t.co/5zHY1SFkbK
You can now connect 20k+ banks from 40+ countries automatically to @actualbudget via @LunchFlow π
This marks the first use of Lunch Flow's public API, and first repo in the Github org! Couldn't imagine a better start with a well-established app & a great community.
It's been an interesting experience exploring terminal UIs and publishing packages via npm, feedback and contributions are welcome!
https://t.co/p1MTsPbFo7
Just redesigned @LunchFlow 's pricing with the help of @figma and @cursor_ai π¨βπ³
Wanted to communicate:
- Multi-currency support
- Savings on annual plans
- Pay per bank connection
- Limited-time annual savings
Feedback welcome π
Curious if others have experience in leveraging LLMs over their internal datasets for customer support. Feedback and insights are very welcome π
#saas#customersupport#buildinpublic
With @LunchFlow growing past 130 paying customers, I am leveraging AI agents to keep the support workload under a few hours a week, while aiming to respond within a few minutes of users reaching out.
3. Perhaps offering an MCP server for end users to integrate their financial data with their LLMs, imagine the possibilities if you could have a conversation with your bank data in Claude or ChatGPT (or consider using your own local LLM, for obvious privacy concerns π)