Pinned Tweet
You can start building without being a developer.
Start with one repeated workflow, then choose the route that fits.
The same Friday watchlist notes can become a prompt brief or a small API map.
From idea to app, with or without code. Review still comes before production.
When an AI agent gets a rate limit error (429) on one watchlist edit, every other watchlist edit on that key is out of budget too. Does your agent wait as long as the 429 says before it tries again?
A chart is a stack of candles. etoro's MCP server lets an AI assistant fetch them, so a small build could draw one from scratch. What would you make from that stack? #AIWednesday
People who sign in to your app with etoro approve what it is allowed to do for them. What would your app ask them to allow first?
Price alerts in. Sheet rows out.
An @n8n_io workflow built against etoro's documented price alert endpoints, shown with stub data. Read and log only, no trading. Placeholder symbols and prices.
Your first build with etoro's developer tools can be just for you, shaped by how you work. Or it can be for other people too, shaped by what they need. Which is yours, and why? #TipsTuesday
An AI assistant writing code for an API it's never seen has to fill in the gaps on its own. etoro made a connector so the assistant can look up etoro's real API instructions first, then write the code. builders.etoro.com/tools/mcp
Sunday with etoro's Claude Code guide: ask it to set up etoro's add-on, spot etoro in its list, then bring it questions about etoro's API.
Which step are you on? Not started counts too.
Apps built on etoro's Public API can set price alerts.
Say you want to change one. Would you rather edit it, or delete it and start fresh, and why?
Review triggers are already half a spec. With the etoro MCP server, an AI coding tool starts from the API docs and schemas.
Building this? Which tools would others use?
Past performance is not an indication of future results. Capital at risk.
nitter.gq/IWAlshe7i/status/21102…
Your app ships. Every test passes. Then a real user hits the case nobody wrote down.
If your app runs on etoro data: what was that case, and what check catches it now before a release goes out?
Paste one prompt.
Get a portfolio history desk.
Builders MCP pulls the live book. The dashboard shows what moved. Suggested checks stay human - review, watch, keep. No auto-send.
Builders: what is the first check your history desk should surface?
NFA.
One live call.
Then the fun starts.
Open the Builders Playground. Run one real request. Paste it into your AI. A builder owns that loop so the tools feel simple from the first hour.
Builders: what first Playground call would you show a beginner?
NFA.
The catalog said ready.
The call still failed.
Listed is not live.
Before the agent plans around a name, prove it can run:
get-all-routes -> get-route-spec -> one real read
No live mark? Drop the name.
Builders: what do you check before you trust a tool is real?
NFA.
Why connect Builders MCP at all?
So Cursor or Claude can call the same portfolio and market tools the dashboard uses, without copy-paste JSON between apps.
NFA.
Bookmark this one.
Interview this portfolio.
Read-only. MCP or connector + Cursor/Claude Code.
Write stays off.
Full prompt in the next reply.
Answer in plain words.
Then get a one-page memo:
what looks concentrated, what feels neglected, what to research next.
Write stays off the whole time.
If a position is invented, throw the answer out.














