mercury24aug

4 ms·

1.
MCP safety is a big concern, so we created BDSMCP, a proposal to make MCP safer(maxockner.com)
1 points by mercury24aug 6mo ago|2 comments
2.
by mercury24aug 6mo ago
MCP is a relatively new protocol, and we are still ironing out some of the kinks. So we created BDSMCP, a proposal for how to make MCP safer for everyone.
3.
by mercury24aug 10mo ago
Giving LLM the ability to generate UI is a cool concept, but our models are not there yet. MCP Apps can be extremely powerful, for example, you can play Doom inside ChatGPT: https://x.com/rauchg/status/197823516139
4.
by mercury24aug 10mo ago
This is a very strict definition of MCP. An agent (with LLM call inside) can be an MCP. Event a UI component can be an MCP.
5.
by mercury24aug 10mo ago
AI fills in a form, and you want to adjust the form before clicking submit. How often do you have to adjust AI's answer vs accepting it as is
6.
MCP Apps: Extending servers with interactive user interfaces(blog.modelcontextprotocol.io)
192 points by mercury24aug 10mo ago|121 comments
7.
by mercury24aug 10mo ago
Looks like OpenAI, Anthropic, and the MCP-UI team actually worked together on a common standard for MCP Apps: https://blog.modelcontextprotocol.io/posts/2025-11-21-mcp-ap... Honestly, I think the biggest friction for M
8.
by mercury24aug 11mo ago
Thank you so much! I'm glad you like it
9.
Show HN: OpenAI hasn't released their Apps SDK so we did(github.com)
12 points by mercury24aug 11mo ago|3 comments
10.
by mercury24aug 11mo ago
It's funny how OpenAI announced Apps SDK without the SDK. Anyway, we was so excited to get my hands dirty that we built our own SDK: https://github.com/fractal-mcp/sdk
11.
by mercury24aug 12y ago
This account was created 4hr ago just to write negative comments on this thread.