Build a Daily Digest Agent with Cursor
Starting a work session with current context is underrated. If you’re building in a fast-moving space, a quick digest on recent developments, what competitors shipped, what the community is talking about, any relevant announcements, is useful framing before writing code. With AgentPatch connected, Cursor can pull live news results and format them however you want.
Why This Matters
The alternative is doing it manually: opening tabs, checking multiple sources, synthesizing the information yourself. That takes 15 to 20 minutes and usually gets skipped when you’re busy.
Cursor with Google News access through AgentPatch can produce a structured briefing in one query. You get the context without the overhead.
This is also useful for teams. You can ask Cursor to generate a digest, save it to a file in your repo, and commit it. A lightweight way to keep a running log of industry developments alongside your code.
Setup
The AgentPatch CLI is designed for AI agents to use via shell access. Install it, and your agent can discover and invoke any tool on the marketplace.
Install (zero dependencies, Python 3.10+):
pip install agentpatch
Set your API key:
export AGENTPATCH_API_KEY=your_api_key
Example commands your agent will use:
ap search "web search"
ap run google-search --input '{"query": "test"}'
Get your API key from the AgentPatch dashboard.
- Open Cursor Settings (Cmd+Shift+J on Mac, Ctrl+Shift+J on Windows/Linux).
- Navigate to the MCP section in the sidebar.
- Click Add new global MCP server.
- Cursor opens
~/.cursor/mcp.json. Add the AgentPatch server config:
{
"mcpServers": {
"agentpatch": {
"url": "https://agentpatch.ai/mcp",
"headers": {
"Authorization": "Bearer YOUR_API_KEY"
}
}
}
}
- Replace
YOUR_API_KEYwith your actual key from the AgentPatch dashboard. - Save and restart Cursor. It discovers all AgentPatch tools automatically.
Example
You open Cursor in the morning before a sprint. In the chat:
“Before we start coding, give me a quick digest. Pull recent news on: ‘TypeScript’, ‘edge computing’, ‘Cloudflare Workers’, and ‘open source AI models’. Format it as a briefing with the top 3 stories per topic and a one-sentence summary for each.”
Cursor calls Google News for each topic through AgentPatch, pulls recent articles, and formats the output:
## TypeScript
1. TypeScript 5.8 released with improved type narrowing — [source]
2. Deno 2.3 ships with full TypeScript-first API — [source]
3. Survey: TypeScript adoption hits 78% among JS developers — [source]
## Edge Computing
1. Fastly announces GPU-backed edge workers — [source]
2. Akamai and Cloudflare trade benchmarks on cold start times — [source]
3. Edge computing market projected at $232B by 2028 — [source]
## Cloudflare Workers
[...]
## Open Source AI Models
[...]
You notice something relevant and follow up:
“That Cloudflare Workers item about cold start benchmarks, pull more coverage on that.”
Cursor runs a targeted query and brings back the full picture. You now have context for a decision about your own infrastructure without having opened a browser.
For a recurring workflow, you can save the prompt as a snippet and run it at the start of each session. Same effort as typing a command, but you get a current briefing every time.
Wrapping Up
One AgentPatch MCP connection gives Cursor access to Google News and the entire tool marketplace: web search, trends, email, image generation, and more. Add it once, get everything. Get started at agentpatch.ai.