Ask a model "what happened in AI this week?" and it answers from memory. Its knowledge stops at a training cut-off, so for anything recent it either refuses or guesses. In our own test, a plain model even described its made-up stories as "hypothetical".
So we built NestDaddy: live web search, real-time news from around the world and full article text for AI agents. It runs on our own search index, so it does not resell another provider's results.
Today we launched it on Product Hunt: https://www.producthunt.com/posts/nestdaddy
The fix: one MCP config
Add this to Cursor (or Claude Desktop, Windsurf and any MCP client that takes a URL):
{
"mcpServers": {
"nestdaddy": {
"url": "https://mcp.nestdaddy.com/mcp",
"headers": { "Authorization": "Bearer YOUR_KEY" }
}
}
}
Claude Code is one line:
claude mcp add --transport http nestdaddy https://mcp.nestdaddy.com/mcp \
--header "Authorization: Bearer YOUR_KEY"
Ask the same question again and the agent searches, reads the sources and answers with links you can check.
Not using MCP? Plain REST
curl -s "https://nestdaddy.com/api/v1/search?q=python+asyncio+guide&limit=5" \
-H "X-API-Key: $NESTDADDY_API_KEY"
curl -s "https://nestdaddy.com/api/v1/news/search?q=interest+rate&country=US&days=7&limit=10" \
-H "X-API-Key: $NESTDADDY_API_KEY"
More examples in Python, JavaScript and curl: nestdaddy/nestdaddy-examples.
Try it
There is a free plan with no card: get a key at nestdaddy.com/mcp. Launch week only, code PH50 gives 50% off any paid plan for 3 months (until Nov 24).
Feedback on result quality is what we want most. If a search gives you a bad or stale result, tell us in the comments here or on the Product Hunt page and we will look at it.
Top comments (0)