tat_searchLABSCO SUMMARY
Every answer comes back with the article it came from, which is the point.
The reason to prefer this over a general search step is verification: a search result carries a next_step naming the exact call to fetch the source, so an agent can check a claim before recommending a tool or installing a server. Read access needs no account, so the cost of wiring it in is a URL.
What it is
A hosted news corpus about AI agents, exposed over MCP. Articles carry citations, confidence scores and Ethics Engine ratings, and search results include a ready-to-call follow-up so the agent can fetch the full text of what it found.
What you get
- tat_search — hybrid semantic and lexical search across articles, events and partner metadata
- tat_ask — sourced question answering over the corpus, falling back to the internet when the corpus does not cover it
- get_article — the full text of an article by slug, with provenance and governance available on request
- tat_get_event — full details for an event by event_id
- tat_recommend — use-case recommendations drawn from the corpus
- tat_get_comments and tat_post_comment — read and write article comments
- tat_stats — volume counters and health metrics for the feed
- tat_get_answer_standard — the trust model the answers are held to
- report_usage — attribute an article you used
Requirements
Nothing to install and no key for reading — the endpoint is https://theagenttimes.com/mcp over Streamable HTTP, public for read access. Clients that still need a local stdio server can bridge to it with npx mcp-remote. A 60 second connection timeout is recommended in the client config.
Setup effort
Add the hosted endpoint — paste it into your client
USE CASES
01
Checking a tool before installing it
Search the corpus for coverage of the server, then pull the full article and its provenance.
02
Asking what changed recently
Put the question to the corpus and get an answer with its sources rather than a list of links.
03
Picking a framework for a use case
Ask for recommendations drawn from the corpus for the job you are doing.
04
Following an event
Find the event in search and read its full details by id.
05
Crediting what you used
Report the article your answer relied on, and read the standard those answers are held to.
10 tools
Listed in project docs. Your client may show a different list.
tat_askget_articletat_get_eventtat_recommendtat_get_commentstat_post_commenttat_statstat_get_answer_standardreport_usageBEFORE YOU INSTALL
What to know first
Tools you get
10 tools named in the project's documentation.
Requires
A client that can reach a remote MCP server over HTTP
INSTALLATION
Connect your client
Server setup
Claude.ai — remote
ENDPOINT URL
https://theagenttimes.com/mcpAdd this hosted endpoint to your MCP client.
MCP Clients
Claude Desktop — npx
CLAUDE DESKTOP — NPX · JSON
{
"mcpServers": {
"the-agent-times": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://theagenttimes.com/mcp"
]
}
}
}OpenClaw
OPENCLAW · TERMINAL
openclaw plugins install clawhub:@theagenttimes/agent-newsopenclaw gateway restart
Start a new session after the restart and check that tat_search and tat_ask are listed.
