Skip to content

gdelt-mcp-server

io.github.cyanheads/gdelt-mcp-servercyanheadsstreamable-httpv0.10.1Scanned

Search and analyze global news coverage and US TV transcripts via the GDELT Project APIs.

11 tools: 11 only read.

Tools: 0 delete, 0 irreversible, 0 write, 11 read.
  • Reads private data · 1
  • Reads text others write · 11
  • Can send data out · 0
  • No findings

Tools listed without sign-in. Remote posture

With Anlyon

Decide which of this server's tools your agent can use.

  • Anlyon lists the tools and adopts none.
  • Each tool you adopt gets approvals and receipts.
  • Your agent never holds the key.
Govern this server →

A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.

What the gdelt-mcp-server can do

The gdelt-mcp-server is published by the GitHub account cyanheads. It has 11 tools. Every one of them only reads: going by their definitions, none of them creates, changes, sends or deletes anything.

1 tool reads private data (gdelt_search_articles); 11 tools bring in text other people can write (gdelt_search_articles, gdelt_search_sentences, gdelt_get_coverage_timeline and 8 more).

None of the fixed checks matched anything in its tool definitions. That is not a safety guarantee.

Its remote endpoint lists its tools without sign-in, and its tools changed 4 times since Anlyon Radar first saw them.

Last scanned Nov 18, 2026, version 0.10.1.

Tools

Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.

11 tools

Only reads 11Show 11 read-only tools
  • gdelt_get_coverage_breakdownBreak down news coverage volume over time by source language or source country, returning a multi-series time series (one series per language or country).Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_get_coverage_breakdown" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Break down news coverage volume over time by source language or source country, returning a multi-series time series (one series per language or country). Shows which countries or languages drove early vs. late coverage — useful for tracing how a story propagated geographically or across language communities. Ranks every series by the sum of its values and returns limit of them per page (default 10, cap 75); when more remain it returns nextOffset — pass it back as offset with otherwise identical inputs for the next page. otherSeriesLabels names every series not on the page — pass any labels as the series input to get only those series, complete. Every series carries sum (its ranking key), pointCount, and peak; includePoints: false leaves out its data points for a lighter ranking. GDELT answers a timestep without matching coverage with 0 and leaves out a timestep its own data lacks; missingTimesteps counts those left out between the first and last returned. Each series whose name maps to one GDELT source code carries operator (sourcecountry:<FIPS> or sourcelang:<language>) — add it to a gdelt_search_articles or gdelt_get_coverage_timeline query to read that source's articles or volume. Values are normalized: each point is the topic's share of that source's media output, not an absolute article count, so values from different sources do not add up to a share. Small media markets with concentrated coverage therefore rank above large markets with diverse output — a high value means the topic dominated that source's coverage, not that it published the most articles. Use breakdownBy "country" with the signal-detection chain to map geographic attention, or "language" to detect non-English media surges.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_get_coverage_timelineRetrieve a time series showing when news coverage of a topic spiked, or how average tone shifted over time.Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_get_coverage_timeline" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Retrieve a time series showing when news coverage of a topic spiked, or how average tone shifted over time. Use mode "volume" for normalized coverage intensity (% of all global coverage per timestep). Use "volume_raw" for distinct matching-article counts per timestep and norm, the unsmoothed monitored-article denominator; smoothed counts are moving averages. Use mode "volume_with_articles" for the same signal plus the top articles that drove each spike — this is the primary signal-detection mode: a single call reveals both the spike and its cause, avoiding a follow-up gdelt_search_articles call. Use mode "tone" for average sentiment score per timestep (negative = hostile/fearful, positive = celebratory). Date resolution is the one GDELT reports, or else inferred from returned intervals: 15 minutes or hours for short windows, days for longer ones. GDELT answers a timestep without matching coverage with 0 and leaves out a timestep its own data lacks; missingTimesteps counts those left out between the first and last returned. In volume_with_articles mode each timestep carries its first 3 article links and referenceCount, the number of references GDELT supplied (up to ten), not all matching articles; name timesteps' dates in points to get only those timesteps, each with its full supplied list. A response carries limit timesteps (default 10, cap 400); when more remain it returns nextOffset — pass it back as offset with otherwise identical inputs for the next page. Explicit startDatetime/endDatetime windows may predate the last 3 months in "volume" and "tone" modes; "volume_raw" and "volume_with_articles" cover only the last 3 months.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_get_tone_distributionGet the tonal distribution of articles matching a query as a histogram (bins approximately -30 to +30).Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_get_tone_distribution" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Get the tonal distribution of articles matching a query as a histogram (bins approximately -30 to +30). Unlike a single average tone score, the histogram reveals whether coverage is uniformly negative, bimodal (some articles extremely positive and some extremely negative), or clustered near neutral. Each bin carries its first 3 representative article links and referenceCount, how many GDELT returned; name bins in the bins input to get only those bins, each with every link. A response carries limit bins (default 10, cap 400); when more remain it returns nextOffset — pass it back as offset with otherwise identical inputs for the next page. Distinct from gdelt_get_coverage_timeline (mode: tone) — this is a snapshot distribution across all matching articles, not a time series. Use gdelt_get_coverage_timeline with mode "tone" to see how sentiment shifted over time.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_get_tv_breakdownRank TV stations or shows mentioning a query.Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_get_tv_breakdown" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Rank TV stations or shows mentioning a query. Station values are percentages of each station’s monitored airtime, or raw matching 15-second clip counts with normalize=false. Show values always represent percentages of matching clips, even with normalize=false. Labels are preserved exactly; ellipsized show labels may not work as exact show: selectors. Returns 25 rows by default, up to 10,000 per page. Archive coverage varies by station; use gdelt_list_tv_stations for station IDs, network and market selectors, and monitoring date ranges.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_get_tv_clipsRetrieve the top matching TV news clips for a query from the Internet Archive's Television News Archive: returns every fetched in-window clip, up to 3,000.Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_get_tv_clips" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • untrusted-inputThe description's first sentence says "retrieve" and "internet": it reads web pages, which anyone can write.
    Description, as the server wrote it
    Retrieve the top matching TV news clips for a query from the Internet Archive's Television News Archive: returns every fetched in-window clip, up to 3,000. Each clip includes show name, the station's display name — with its station ID when that name resolves to one — air timestamp, a 15-second transcript excerpt, and a direct link to view the full one-minute clip. Use after gdelt_search_tv to read the actual transcript content driving a coverage spike — pass a series' station value as stations here. GDELT answers a TV window in whole clock hours when it spans up to 7 days, whole UTC days when shorter than 3 years, and whole UTC months beyond, so clips it returns from outside an explicit startDatetime/endDatetime window are dropped — though they still spend maxRecords. 3,000 is a hard per-call ceiling and GDELT offers no cursor: when a query fills it, re-query narrower startDatetime/endDatetime windows — the response hands back the exact windows to use. Archive coverage varies by station; use gdelt_list_tv_stations for station IDs and monitoring date ranges.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_get_tv_contextGet the top co-occurring words and phrases from TV news clips matching a query — the vocabulary framing a topic on television.Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_get_tv_context" says "get": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Get the top co-occurring words and phrases from TV news clips matching a query — the vocabulary framing a topic on television. Returns the most frequent non-stopword terms with relative frequency scores (0–100, where 100 = the query term itself). Terms and scores come from GDELT's relevance-ranked sample of matching clips, documented as the 200 most relevant, not all matching coverage — compare windows or stations with a separate, narrower call for each. Use to understand narrative framing, identify related concepts mentioned alongside a topic, or generate follow-up search terms. Returns 25 terms by default, with count pages up to 25,000 terms. Archive coverage varies by station; check gdelt_list_tv_stations for each station’s monitoring date range.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_list_tv_stationsList the television stations available for TV search with their market, network, monitoring start date, and monitoring end date — every station by default, or only those matching t…Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_list_tv_stations" says "list": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    List the television stations available for TV search with their market, network, monitoring start date, and monitoring end date — every station by default, or only those matching the optional stations, network, and market filters (each an exact, case-insensitive match; combined with AND). activeCount and totalCount count the returned stations. Stations with an end date within the last 24 hours are flagged as active; stations with earlier end dates are discontinued. Use before querying to verify a station was active during the target time period, or to discover valid station IDs for the stations parameter in other TV tools. Archive coverage varies by station; check each station’s monitoring date range. Use the returned network values with network: and market values with quoted market: selectors in TV queries.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_search_articlesSearch the last 3 months of global news coverage (65+ languages) using the GDELT DOC API.Reads private dataReads text others writeNo findings

    Why it is in this group

    • The name "gdelt_search_articles" says "search": a word for reading.

    What it can reach

    • private-dataThe description's first sentence says "search" and "doc": it reads documents.
    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Search the last 3 months of global news coverage (65+ languages) using the GDELT DOC API. Fetches up to 250 articles with URL, title, source domain, language, country, the date GDELT saw each one, and social image URL, and returns every fetched article inside the requested window. Query supports full GDELT syntax: phrases ("bird flu"), boolean OR ((flu OR pandemic)), source country (sourcecountry:china), source language (sourcelang:spanish), domain (domain:who.int), GKG theme (theme:TAX_DISEASE_OUTBREAK — find identifiers with gdelt_search_themes), tone filter (tone<-5 for negative), proximity (near20:"flu virus"), and repeat (repeat3:"outbreak"). 250 is a hard per-call ceiling and GDELT offers no cursor: when a query fills it, re-query narrower startDatetime/endDatetime windows — the response hands back the exact windows to use. Note: this API covers only the most recent 3 months — use gdelt_search_tv for historical TV transcripts back to 2009.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_search_sentencesSearch the sentences of the last 72 hours of English-language news coverage with the GDELT Context 2.0 API, to read what coverage actually said.Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_search_sentences" says "search": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Search the sentences of the last 72 hours of English-language news coverage with the GDELT Context 2.0 API, to read what coverage actually said. Every query term must fall inside one sentence; each match returns that sentence, the text around it, and its article URL, title, domain, and seen date — one sentence per article. Matches come newest first (dateDesc) or oldest first (dateAsc); there is no relevance ranking. maxRecords counts matching sentences before GDELT keeps one per article, so pages run short, and a page that is not every match hands back continuationWindow, the startDatetime/endDatetime that reaches the next matches. GDELT reads window bounds only to the minute, so a page whose matches all fall in one minute cannot be continued inside it: raise maxRecords, and at its 200 ceiling continuationWindow skips past that minute. Query operators: phrases ("bird flu"), OR groups ((ukraine OR gaza) ceasefire), -term to exclude an exact word, near5:"trump putin", and repeat2:"ukraine"; quotesOnly keeps sentences holding a quotation mark. For the DOC-only filters (sourcecountry:, sourcelang:, theme:, tone), domain filtering, other languages, or coverage older than 72 hours, use gdelt_search_articles.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_search_themesFind GDELT Global Knowledge Graph (GKG) theme identifiers for the theme: operator that gdelt_search_articles, gdelt_get_coverage_timeline, gdelt_get_tone_distribution, and gdelt_ge…Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_search_themes" says "search": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    Description, as the server wrote it
    Find GDELT Global Knowledge Graph (GKG) theme identifiers for the theme: operator that gdelt_search_articles, gdelt_get_coverage_timeline, gdelt_get_tone_distribution, and gdelt_get_coverage_breakdown accept in query. Searches the identifiers in the GDELT GKG theme lookup, which carries no labels or descriptions: every query word must begin one of the _-separated parts of an identifier or run across consecutive parts, or all the words joined must, so "drought" finds NATURAL_DISASTER_DROUGHT, "cyberattack" finds CYBER_ATTACK, "plant disease" finds TAX_PLANTDISEASE, and "wb water" narrows to World Bank water themes. There is no stemming or synonym matching — "displacement" does not reach DISPLACED — except one fallback: when nothing matches, the search retries once with a trailing s dropped from each word of four or more letters, and says so. Matches rank an exact identifier first, then by the count the lookup lists — a static prevalence figure, not a live article total — and each carries a paste-ready operator such as theme:TAX_DISEASE_OUTBREAK. Page with offset and limit (up to 100 matches per page; a larger limit runs at 100).

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true
  • gdelt_search_tvSearch television news closed captions for spoken mentions of a query.Reads text others writeNo findings

    Why it is in this group

    • The name "gdelt_search_tv" says "search": a word for reading.

    What it can reach

    • untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
    • untrusted-inputThe description's first sentence says "search" and "mention": it reads comments, which other people can write.
    Description, as the server wrote it
    Search television news closed captions for spoken mentions of a query. Returns a bounded, paged per-station time series showing airtime devoted to the topic — 500 points by default, up to 20,000 per page. Each series carries GDELT's display name for the station and, when that name resolves to one, the station ID gdelt_get_tv_clips takes as stations. Use stations to select station IDs (e.g. ["CNN", "FOXNEWS", "MSNBC"]), or select with station:, network:, or quoted market: operators in query. TV query also supports in-query operators: station:CNN, network:CBS, market:"National", show:"Anderson Cooper 360", context:"vaccine". Archive coverage varies by station; use gdelt_list_tv_stations for station IDs, network and market values, and monitoring date ranges.

    Annotations the server declares

    • readOnlyHint: true
    • openWorldHint: true

This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.

Exfil surface

Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.

No private-data, untrusted-input, outbound path in this server alone: no tool is tagged outbound.

Reads private data1
Reads files, mail, messages, databases, secrets, calendars or repositories.
Can send data out0
Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.

Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.

Remote posture

What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.

Tools listed without sign-in

Looked at Nov 20, 2026, 19:31 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.

Reachable
yes
HTTP status of an anonymous initialize
200
tools/list served without credentials
yes
Tools that change things, listed without credentials
0
Protocol version
2025-06-18
WWW-Authenticate on 401
not observed
resource_metadata in WWW-Authenticate
not observed
RFC 9728 protected resource metadata
not observed
Authorization servers
not observed
Legacy HTTP+SSE transport
no
Dynamic client registration
not observed
Client ID metadata documents
not observed

Findings

Matches of fixed checks, with the text they matched.

No check matched anything in these tool definitions. That is not a safety guarantee.

History

Changes to its tool definitions and to its remote endpoint, newest first.

  1. Tools7 changed.~7 changedView diff
  2. Tools1 added, 8 changed.+1 added~8 changedView diff
  3. Tools7 changed.~7 changedView diff
  4. Tools1 added, 9 changed.+1 added~9 changedView diff

Install

Pinned to version 0.10.1. Values in angle brackets are yours to fill in.

claude mcp add --transport http gdelt https://gdelt.caseyjhand.com/mcp
claude mcp add --transport http gdelt https://gdelt.caseyjhand.com/mcp

Run in a terminal.

Claude Code permissions

Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.

.claude/settings.json
{
  "permissions": {
    "allow": [
      "mcp__gdelt__gdelt_search_articles",
      "mcp__gdelt__gdelt_search_sentences",
      "mcp__gdelt__gdelt_get_coverage_timeline",
      "mcp__gdelt__gdelt_get_tone_distribution",
      "mcp__gdelt__gdelt_get_coverage_breakdown",
      "mcp__gdelt__gdelt_search_themes",
      "mcp__gdelt__gdelt_search_tv",
      "mcp__gdelt__gdelt_get_tv_breakdown",
      "mcp__gdelt__gdelt_get_tv_clips",
      "mcp__gdelt__gdelt_get_tv_context",
      "mcp__gdelt__gdelt_list_tv_stations"
    ],
    "ask": [],
    "deny": []
  }
}

Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as gdelt. Review it before you use it.

Capability tags behind 11 rules
  • mcp__gdelt__gdelt_search_articles private-data, untrusted-input
  • mcp__gdelt__gdelt_search_sentences untrusted-input
  • mcp__gdelt__gdelt_get_coverage_timeline untrusted-input
  • mcp__gdelt__gdelt_get_tone_distribution untrusted-input
  • mcp__gdelt__gdelt_get_coverage_breakdown untrusted-input
  • mcp__gdelt__gdelt_search_themes untrusted-input
  • mcp__gdelt__gdelt_search_tv untrusted-input
  • mcp__gdelt__gdelt_get_tv_breakdown untrusted-input
  • mcp__gdelt__gdelt_get_tv_clips untrusted-input
  • mcp__gdelt__gdelt_get_tv_context untrusted-input
  • mcp__gdelt__gdelt_list_tv_stations untrusted-input

Pin

Keep the decision: fail CI when the tools change, follow every change, or show a badge.

Pin in CI

Fail a build when the tools you reviewed are not the tools the server serves today.

curl -s "https://radar.anlyon.com/api/v1/servers/io.github.cyanheads/gdelt-mcp-server/verify?hash=409564df4a4824906fbf837ce1752b83328e28ccb984bf25700fe954f3342f3a" | jq -e '.changed == false'

Pinned hash 409564df4a48, from Nov 18, 2026, 08:18 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://radar.anlyon.com/servers/io.github.cyanheads/gdelt-mcp-server/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

Anlyon Radar badge for io.github.cyanheads/gdelt-mcp-server
[![Anlyon Radar](https://radar.anlyon.com/badge/io.github.cyanheads/gdelt-mcp-server.svg)](https://radar.anlyon.com/servers/io.github.cyanheads/gdelt-mcp-server)

Source: MCP Registry · repository. Synced Nov 18, 2026, 08:17 UTC.

gdelt-mcp-server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the gdelt-mcp-server safe?

No scan can promise that. What its definitions show: 11 tools, all of which only read, and 0 findings from fixed checks. Read-only tools can still pass private data to a model.

Can the gdelt-mcp-server delete data?

No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.

Can the gdelt-mcp-server do things that cannot be undone?

No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.

What tools does the gdelt-mcp-server have?

11 tools: 11 read. gdelt_search_articles, gdelt_search_sentences, gdelt_get_coverage_timeline, gdelt_get_tone_distribution, gdelt_get_coverage_breakdown, gdelt_search_themes, gdelt_search_tv, gdelt_get_tv_breakdown and 3 more.

Does the gdelt-mcp-server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the gdelt-mcp-server send data to outside services?

None of its tools sends, posts or publishes, or takes a recipient, channel or URL.

Has the gdelt-mcp-server changed its tools?

Yes, 4 times since Anlyon Radar started tracking it, most recently on Nov 18, 2026. Every change is on this page and in its Atom feed.