arxiv_recentLABSCO SUMMARY
Monday morning's reading, assembled in one call.
Most trend servers cover a single source. This one covers seven and treats them as different: arXiv gets per-category round-robin so a small category still surfaces, FDA recalls get class filters, GitHub search gets a date floor. The briefing tool is what turns that into a habit rather than a query.
What it is
A single MCP server pulling trend data from seven sources — arXiv, PubMed, Hugging Face Daily Papers, GitHub trending and search, Hugging Face models and datasets, openFDA 510(k) clearances and openFDA recalls — and rendering newspaper-style briefings, with per-domain tuning configured once.
What you get
- Recent and searched arXiv papers, with per-category round-robin so small categories are not drowned by large ones
- PubMed search returning full abstracts
- Hugging Face Daily Papers sorted by community upvotes
- GitHub trending, and a search filtered by creation date
- Hugging Face models, datasets and spaces, trending or recent
- openFDA 510(k) device clearances, and recall events with class filtering
- Multi-source digests and briefings that call the sources in parallel
- Source text translated into your conversation language, with identifiers, URLs and metric values preserved verbatim
Requirements
Python, and a one-time configuration run to set the categories and sources for your domain. No tokens required — all seven sources work anonymously, and a token only raises that source's rate limit ceiling.
Setup effort
Build from source — git clone https://github.com/salwks/mcp-techTrend.git
USE CASES
01
Starting the week knowing what moved in your field
One briefing across every source you enabled, formatted to read rather than to parse, and translated into the language you are working in.
02
Following the papers without drowning in one category
Recent arXiv by category, with the round-robin doing the balancing, and a direct search when you know what you are after.
03
Tracking a clinical or medical-device question
PubMed for the literature with full abstracts, 510(k) for what has just been cleared, and recalls for what has just gone wrong.
04
Seeing which repositories and models are taking off
GitHub trending and Hugging Face trending together, with the day's upvoted papers alongside them for the research behind the code.
05
Finding projects started since a date you care about
Search GitHub with a creation-date floor, so what comes back is new work rather than long-established repositories.
16 tools
Listed in project docs. Your client may show a different list.
arxiv_searchpubmed_searchpaperswithcode_trendinggithub_trendinggithub_searchhuggingface_trendingfda_510k_recentfda_recalls_recenttrends_digesttrends_briefingtrends_get_configtrends_set_enabled_sourcestrends_set_arxiv_categoriestrends_set_pubmed_querytrends_set_tokenBEFORE YOU INSTALL
What to know first
Tools you get
16 tools named in the project's documentation. TRENDS_ENABLED_SOURCES controls which source tools register; leaving out the FDA and PubMed sources takes 16 down to 13. trends_digest, trends_briefing and the five configuration tools always register.
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
You clone this one yourself.
Server setup
TERMINAL
git clone https://github.com/salwks/mcp-techTrend.git
cd mcp-techTrend
python3 -m venv .venv
.venv/bin/pip install -r requirements.txtEdit ~/Library/Application Support/Claude/claude_desktop_config.json and add the mcpServers block, pointing command at .venv/bin/python and args at run.py
Restart Claude Desktop
Optionally configure sources/categories/tokens with the interactive TUI:
TERMINAL
python configure.pyMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"trends": {
"command": "<YOUR_PATH_TO_MCP_TECH_TREND>/.venv/bin/python",
"args": [
"<YOUR_PATH_TO_MCP_TECH_TREND>/run.py"
]
}
}
}
