Skip to main content
Glama
NewsCatcher MCP

NewsCatcher MCP

Official

Start a research run

start_research

Build verified news datasets or company news tables with source links; start a research run, then follow its run_id. Check time and cost first with a free dry run.

Instructions

Start the NewsCatcher research agent: it searches the news, checks each event against its sources and returns a verified table (one row per event, with source links). Runs take minutes and cost credits, so use it for datasets, not quick lookups (use search_news for those). Two kinds: (1) events: give a query describing what happened, e.g. "layoffs at US tech companies in October 2026" → rows like company, date, headcount, source_url; columns are chosen for the question. (2) companies: give companies (domains) or a company_list_id → every recent story about each company (funding, M&A, leadership, product, legal…), optionally filtered by query. keep_updated turns it into a monitor. dry_run=true checks the query and shows time and cost first, free. Returns a run_id; follow it with get_research.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "start_research"
  }
}

Input Schema

NameRequiredDescriptionDefault
modeNoEvents only. fast (default): one agent, 3–10 minutes, cheapest. standard: deep research, each row checked against its sources, with a cited briefing (10–30 min). max: strongest models, highest recall and cost.fast
queryNoEvents: the dataset you want in plain words, with the time window, e.g. "Bank M&A deals announced in the US since October 1, 2026" or "Series A rounds by fintech startups in Europe this month". With companies: optional filter on their news, e.g. "funding, acquisitions and leadership changes" (blank = every story).
dry_runNoCheck the request without starting it: what would run, how long it takes, the list-price cost, your credit balance and tips on the query. Free; nothing is started or charged.
max_rowsNoEvents only: rows to return at most (default 50; plans cap it, e.g. 100). Charged per 50 rows past the first 50.
companiesNoResearch these companies' news instead of an events query: domains like ["stripe.com","wise.com"] (up to 50 per run on self-serve plans).
list_nameNoName for the saved list.
keep_updatedNoTurn the run into a monitor once it succeeds: true for the default schedule (daily 09:00 UTC; weekly for standard/max events), or a schedule object. The monitor then delivers only what is new.
save_as_listNoAlso save the typed companies as a reusable list (its id comes back).
wait_secondsNoWait up to this many seconds for the run to finish and return its results (default 0: return the run_id at once; then call get_research).
lookback_daysNoCompanies only: days of news to cover (default 30).
company_list_idNoA saved company list (from save_company_list) instead of typing companies.
include_market_activityNoCompanies only: keep stock moves and analyst ratings (default true).

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A4.6/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations declare readOnlyHint=false, idempotentHint=false and openWorldHint=true, but the description goes well past them: it discloses latency ("runs take minutes"), cost ("cost credits", per-50-row charging is echoed in schema), that dry_run is free and starts nothing, and that keep_updated converts the run into a monitor that "delivers only what is new". Nothing contradicts the annotations.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the purpose and the search_news exclusion, and every sentence carries content. It is a single dense block with heavy parentheticals, though, so the events/companies distinction and the dry_run advice are harder to scan than they need to be.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 12-parameter, all-optional, non-idempotent job-starting tool with no output schema, the description covers what an agent must know: inputs come in two shapes, the call is asynchronous, a run_id must be followed by get_research, and cost/time can be previewed for free. The remaining detail (mode tiers, lookback, schedules) is fully documented in the schema.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so the baseline is 3. The description does give a concrete example query illustrating the `query` parameter and explains the dry_run/keep_updated/return-id flow, but it largely restates what the schema already documents and never mentions the `mode` enum (fast/standard/max), so it adds only marginal meaning beyond the structured fields.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description uses a specific verb+resource ("Start the NewsCatcher research agent") and immediately defines the output shape: a verified table with one row per event and source links. It then splits into the two distinct invocation modes (events vs. companies) and names the sibling it is not (search_news), so an agent can distinguish it without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It states when to use this tool ("datasets, not quick lookups") and the explicit alternative for the other case ("use search_news for those"). It also gives the safe pre-flight path ("dry_run=true checks the query and shows time and cost first, free") and the required follow-up step ("Returns a run_id; follow it with get_research").

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.