Blog
Learn how to scrape the web and create tools with public data.
A Company Monitoring Agent That Doesn't Cry Wolf
The hard part of a news monitoring agent isn't finding coverage. It's deciding which of forty articles is worth interrupting someone for.
MCP Resources vs Tools: Which One Should Search Be?
The protocol gives you two ways to expose data. Picking the wrong one produces a server that technically works and nobody uses well.
Adding Live Search to a Flowise Chatflow
Flowise's Custom Tool node takes plain JavaScript. That's all you need to give a visual chatflow real web results.
A Market Sizing Agent That Shows Its Arithmetic
Ask a model for a market size and you get a number with no provenance. Ask it to build one from sourced inputs and you get something you can argue with.
One Question, Six Endpoints, One Round Trip
Some questions need web results, news, and local listings at once. Fanning out across endpoints costs the latency of the slowest one, not the sum.
Streaming What Your Research Agent Is Doing, Not Just What It Says
Twenty seconds of blank screen while an agent searches feels broken. Streaming the tool calls turns dead time into visible progress.
Reading Answer Boxes and Knowledge Panels Without Trusting Them Blindly
Google's answer box is a fast path to a fact and an occasional confident lie. Use it as a hint, verify it as a claim.
Research Workflows That Survive a Deploy
A twenty-minute research job shouldn't restart from scratch because a pod got rescheduled. Durable execution makes each search call replayable.
Adding a Custom Search Tool to a Google ADK Agent
ADK gives agents a built-in search, but it's tied to one provider. A plain Python function tool gets you control over the whole retrieval layer.
Retry Logic That Doesn't Make Things Worse
The naive retry loop turns a brief rate limit into a sustained one. Getting backoff right takes about forty lines.