io.github.almaestilo/newswake

Official2 tools

by almaestilo · JavaScript

News momentum + what-changed outcome tools for agents. Signals + links, ~296 sources.

io.github.almaestilo/newswake (MCP Server)

The io.github.almaestilo/newswake MCP server provides “news momentum” and “what changed” outcome tools for AI agents, returning decided answers about how news is moving rather than a list of articles. It is described as using signals and links, with support for roughly 296 sources and 2 tools.

🛠️ Key Features

News momentum and “what changed” outcome tools for agents
Returns decided answers (not a pile of articles)
Signals + links
Live service: https://newswake-ajq8t.ondigitalocean.app

🚀 Use Cases

Recurring agent checks that need “what’s new” compared to prior context
Determining how a topic’s status is changing over time

⚡ Developer Benefits

MCP endpoint: https://newswake-ajq8t.ondigitalocean.app/mcp
Uses streamable-http
No key in beta
15 calls/day free

⚠️ Limitations

Beta mode: “no key in beta” implies access constraints are tied to the beta state
Free tier rate limited to 15 calls/day

Topics

ai-agentsmcpmcp-servermodel-context-protocolmonitoringnewsnews-api

Captured live from the server via tools/list.

find_trending

What news is ACCELERATING right now — a ranked list of story-events climbing in attention, optionally scoped to a beat/topic. Each returns trend, momentum, acceleration, outlet count, a runway hint (act now vs window closing), why-now, and a link. Use to decide what to ride/publish/watch. ~2h behind live, ~296 mainstream sources, signals + links only.

Parameters2

topic	string	optional	Beat or keyword to scope to, e.g. "technology", "crypto", "nvidia". Omit for all beats.
limit	integer	optional	Max results (default 8).

Raw schema

{
  "type": "object",
  "properties": {
    "topic": {
      "type": "string",
      "description": "Beat or keyword to scope to, e.g. \"technology\", \"crypto\", \"nvidia\". Omit for all beats."
    },
    "limit": {
      "type": "integer",
      "minimum": 1,
      "maximum": 15,
      "description": "Max results (default 8)."
    }
  },
  "additionalProperties": false,
  "$schema": "http://json-schema.org/draft-07/schema#"
}

whats_changed

The DELTA on an entity — what is new or has shifted in news coverage of a company/person/topic since a given time. Returns new vs updated story-events, trend, importance, and links. This is the stateful "what changed since my last check" that stateless search tools cannot answer. ~296 mainstream sources, signals + links only.

Parameters2

entity	string	required	Company, person, or topic, e.g. "Nvidia", "OpenAI", "Donald Trump".
since	string	optional	ISO timestamp of your last check. Default: 7 days ago.

Raw schema

{
  "type": "object",
  "properties": {
    "entity": {
      "type": "string",
      "description": "Company, person, or topic, e.g. \"Nvidia\", \"OpenAI\", \"Donald Trump\"."
    },
    "since": {
      "type": "string",
      "description": "ISO timestamp of your last check. Default: 7 days ago."
    }
  },
  "required": [
    "entity"
  ],
  "additionalProperties": false,
  "$schema": "http://json-schema.org/draft-07/schema#"
}

README.md2 KBView on GitHub

newswake

News momentum & "what changed" — outcome tools for AI agents. An MCP server that returns decided answers about how news is moving, not a pile of articles.

Live: https://newswake-ajq8t.ondigitalocean.app · MCP endpoint: https://newswake-ajq8t.ondigitalocean.app/mcp (streamable-http, no key in beta, 15 calls/day free).

Why

Search APIs (Exa, Tavily, Brave, Perplexity) are stateless — every call re-searches from zero and remembers nothing about what a topic looked like last week. So an agent doing a recurring check on a company has no clean way to answer "what's new since last time," and the LLM can't reliably fake it. newswake tracks persistent story-events hour by hour, so "momentum" and "what changed" are real time-series.

Tools

find_trending(topic?, limit?) — story-events accelerating in attention right now: trend, momentum, a runway hint (ride now vs window closing), why-now, and a link. Scope to any beat.
whats_changed(entity, since?) — the delta on a company/person/topic: new vs updated story-events since a timestamp, with links. The stateful answer a search API can't give.

Both return structured JSON with confidence + links to original reporting — never article text.

Use it

json

{ "mcpServers": { "newswake": { "type": "http", "url": "https://newswake-ajq8t.ondigitalocean.app/mcp" } } }

Claude Code: claude mcp add --transport http newswake https://newswake-ajq8t.ondigitalocean.app/mcp

Honest scope

~296 mainstream sources, US/English, ~2h behind live (built for monitoring/research loops, not sub-minute reactions). Beta; no SLA. Signals + links only.

Install

Tools (2)

find_trendingwhats_changed

Remote endpoint

Streamable HTTP

Hosted server - connect over the network, no local install.

url

https://newswake-ajq8t.ondigitalocean.app/mcp

Add to VS Code Add to Cursor

claude_desktop_config.json