News search for AI agents: recent news on any topic from outlets worldwide, by country and language.
Scored across 4 tools
Each tool has a clearly distinct purpose: search_news finds articles, get_contents fetches article details, find_similar discovers related articles, and create_chart generates charts. There is no overlap or ambiguity between any pair.
All tool names follow a consistent verb_noun pattern (create_chart, find_similar, get_contents, search_news). While 'find_similar' uses an adjective rather than a noun, the pattern is uniform and predictable.
Four tools is well-scoped for a news search and analysis server. Each tool earns its place, covering search, content retrieval, similarity lookup, and visualization without redundancy or bloat.
The tool surface covers the core lifecycle of news lookup and analysis: search, retrieve, compare, and visualize. There are no obvious missing operations for the stated purpose; even the intentional limitation on full-text access is declared clearly.
Turn a question into a chart card, without thinking about chart types: the standings of the latest race, poll or election results by candidate, wins in a season, coverage over time, share of voice between names, tone, the outlets that cover a topic most, what an outlet published, a timeline of a story, the figures the articles publish (a price, a rate), or data from a custom index. It works out what the user expects (understood, with other readings to pass as interpretation), verifies every value against its source, picks the chart, the key figures and a title that states the finding, and returns an image URL (show it to the user) and an embed URL. If the exact data is not published it shows the closest thing and says so (rung, warnings). It costs its search (the mode it needs, per query) plus US$0.50 per 1,000 charts; the same request within 10 minutes is free.
| Name | Required | Description | Default |
|---|---|---|---|
| days | No | Only the last N days, 1 to 365. By default it depends on the chart: 30 for coverage, 7 for figures. | |
| type | No | auto (default) picks the chart that tells the data best. Or force one: line, area, bar, bar_horizontal, stacked_bar, pie, donut, scatter, funnel, timeline, kpi, table. | auto |
| index | No | Optional: chart your own custom index (idx_…) instead of the news: prices, ratings, availability, brands. | |
| query | Yes | What to chart, in plain words and any language: "blue dollar this week", "Milei vs Bullrich coverage", "tone of the IMF coverage", "which outlets cover lithium", or "prices" with index. | |
| theme | No | light (default), dark, editorial (warm paper) or electric (deep blue). | light |
| compare | No | What to compare, 2 to 5, if the query does not say it with "vs": ["Milei", "Bullrich"]. | |
| countries | No | Only sources from these countries: ISO 3166-1 alpha-2 codes, such as ["AR"] or ["US", "GB"]. | |
| languages | No | Only sources that publish in these languages: ISO 639-1 codes, such as ["es"] or ["en", "pt"]. | |
| interpretation | No | Optional: one of the `alternatives` of a previous call, to chart that reading of the query instead. |
| Name | Required | Description |
|---|---|---|
| id | No | |
| data | No | The values drawn (up to 40 per series). |
| rung | No | exact (the data asked for), partial (a verified part), related, coverage or stories (the closest thing found). |
| type | Yes | |
| table | No | The rows of a table chart (up to 10), keyed by column label. |
| title | Yes | It states the finding. |
| cached | No | |
| events | No | |
| sources | No | |
| cost_usd | Yes | |
| subtitle | No | |
| warnings | No | |
| embed_url | No | An iframe for a web page or app. |
| image_url | No | PNG: show it to the user. |
| improving | No | Still looking for the exact data: the image and embed update themselves if it arrives. |
| request_id | Yes | |
| understood | No | What we understood the user expects to see. |
| key_figures | No | |
| alternatives | No | Other readings of the query: pass one as `interpretation` to chart it instead. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description discloses non-obvious behavior well beyond the annotations: it interprets the user's expectation, verifies values against sources, falls back to the closest data with warnings, returns image and embed URLs, and explains the cost model including a 10-minute free cache. This is substantial added context.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
The description is long but information-dense. It is front-loaded with the core purpose, then explains the internal process, fallback behavior, and cost model. Every sentence contributes useful operational context without filler.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a 9-parameter tool with full schema coverage and an output schema, the description covers the non-obvious aspects: interpretation flow, verification, fallback behavior, output URLs, and pricing. Nothing critical for correct invocation is missing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
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 adds meaning beyond the schema by explaining the interpretation parameter ('other readings to pass as interpretation'), the index parameter ('data from a custom index'), and the auto behavior of the type parameter. It does not need to repeat schema details.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific verb and resource: 'Turn a question into a chart card.' It clearly differentiates from siblings like search_news or get_contents by focusing on chart creation rather than retrieval or content access.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description gives clear context for when to use the tool: whenever a user wants a chart from a question. It does not explicitly name sibling tools or state when not to use it, so it falls short of a 5, but the intended use is unambiguous.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Find other news articles about the same story as a given article URL, across the index (the last 7 days by default). Useful to see how other outlets covered a story. US$2.20 per 1,000 calls.
| Name | Required | Description | Default |
|---|---|---|---|
| url | Yes | The article URL whose story to find elsewhere. | |
| days | No | Only the last N days, 1 to 365. Defaults to 7. | |
| max_results | No | How many results, 1 to 25. Defaults to 10. |
| Name | Required | Description |
|---|---|---|
| mode | Yes | |
| cached | No | |
| results | Yes | |
| cost_usd | Yes | |
| warnings | No | |
| reference | No | |
| incomplete | No | The time or token budget ran out: there may be more. |
| request_id | Yes | |
| near_misses | No | Only when nothing matched: the closest articles, which may not be about it. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true and openWorldHint=true, so the safety profile is covered. The description adds the default 7-day window and pricing, but does not go beyond that into behavior like result ordering, failure modes, or how 'similar' is determined.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three short sentences with no filler; the core action is front-loaded, followed by a use case and pricing. Every sentence earns its place.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple read-only tool with a complete input schema and an output schema, the description covers the essential context: what it finds, the default window, and cost. Nothing critical is missing for an agent to invoke it correctly.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the schema already documents url, days, and max_results. The description only restates the URL anchor and default time window, adding no meaning beyond the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific verb and resource: 'Find other news articles about the same story as a given article URL.' This clearly distinguishes the tool from siblings like search_news (query-based) and get_contents (fetching an article's content) by anchoring on a URL and story similarity.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
It gives a concrete use case ('Useful to see how other outlets covered a story') and sets the default scope ('last 7 days by default'). It does not explicitly name alternatives or exclusion criteria, but the URL-based similarity framing makes the intended context clear.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Get the title, standfirst, date, source and a short verbatim excerpt (up to 25 words) of up to 10 news article URLs. With a query, the excerpt is the one about it and relevance says how much the article covers it. Never returns the full text. US$0.20 per 1,000 pages read (US$0.40 with a query); pages that fail are free.
| Name | Required | Description | Default |
|---|---|---|---|
| urls | Yes | Up to 10 article URLs, such as https://example.com/news/article. | |
| query | No | Optional: the excerpt is then the one about this query, with a relevance score. |
| Name | Required | Description |
|---|---|---|
| results | Yes | |
| cost_usd | Yes | |
| request_id | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Beyond the readOnlyHint annotation, the description adds significant behavioral context: it explicitly says the tool never returns full text, caps excerpts at 25 words, charges per 1,000 pages read, and makes failed pages free. These are non-obvious constraints an agent needs to know.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three sentences with no redundant wording. The most decision-relevant facts are front-loaded: what is returned, the URL limit, the no-full-text guarantee, and pricing. Every sentence earns its place.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
An output schema exists, so return value details are already covered. The description supplies the remaining operational context: URL count limits, excerpt length, pricing, failure handling, and the effect of the optional query. Nothing essential for invoking the tool correctly is missing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
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 largely restates the schema: urls are 'up to 10 news article URLs' and query is 'the excerpt is then the one about it.' It adds a small clarification that relevance measures 'how much the article covers it,' but does not substantially go beyond the schema.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a precise verb and resource: 'Get the title, standfirst, date, source and a short verbatim excerpt...' and enumerates exact output fields. It does not explicitly differentiate from sibling tools like search_news or find_similar, but the URL-list input and 'Never returns the full text' make its scope clear.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description gives no guidance on when to use this tool versus sibling tools such as check_coverage, find_similar, or search_news. The 'With a query' clause explains optional behavior but does not state when to prefer this tool or what alternatives exist.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Search recent news on any topic across a curated index of news outlets worldwide, judged by a relevance model. Returns the matching articles: title, link, source, date, country and language, standfirst, and short excerpts in the modes that read. Use it for current events and for what outlets reported about a company, person, place or topic. Defaults: mode "fast" (US$1.40 per 1,000 searches) and the last 7 days; mode "ultra" judges headlines only, for US$1.00.
| Name | Required | Description | Default |
|---|---|---|---|
| days | No | Only the last N days, 1 to 365. Defaults to 7 unless published_after or published_before are given. | |
| mode | No | fast (default): headlines and standfirsts · ultra (cheapest): headlines only · normal: also reads the best matches · deep: reads more, finds the topic in other words too. | fast |
| index | No | Optional: search your own custom index (its id, idx_…) instead of the news index: your own list of URLs, sitemap or feed, created at https://app.typesearch.ai/indexes or with POST /v1/indexes. Same modes and prices; no date filter unless you pass days or dates. | |
| query | Yes | What to look for, in any language: a topic, event, person, company or place, such as "inflation in Argentina" or "OpenAI funding". | |
| countries | No | Only sources from these countries: ISO 3166-1 alpha-2 codes, such as ["AR"] or ["US", "GB"]. | |
| languages | No | Only sources that publish in these languages: ISO 639-1 codes, such as ["es"] or ["en", "pt"]. | |
| max_results | No | How many results, 1 to 25. Defaults to 10. | |
| exclude_domains | No | Never these domains or paths. | |
| include_domains | No | Only these domains or paths, such as example.com or example.com/sports. | |
| published_after | No | Published on or after this date: 2026-09-25, or a date-time with offset. | |
| published_before | No | Published on or before this date; a bare date includes that whole day. |
| Name | Required | Description |
|---|---|---|
| mode | Yes | |
| query | Yes | |
| cached | No | |
| results | Yes | |
| cost_usd | Yes | |
| warnings | No | |
| incomplete | No | The time or token budget ran out: there may be more. |
| request_id | Yes | |
| near_misses | No | Only when nothing matched: the closest articles, which may not be about it. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With readOnlyHint=true and openWorldHint=true already in annotations, the description still adds meaningful behavior context: exact pricing per mode (US$1.40 vs US$1.00 per 1,000 searches), the default 7-day window, and the fact that some modes 'read' article bodies while others judge headlines only. No contradiction with the read-only/open-world hints.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Three sentences with zero filler: purpose first, then return fields and use cases, then defaults and pricing. Every sentence earns its place, and the most critical scoping info ('any topic', 'curated index', 'recent') is front-loaded.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For an 11-parameter tool this is well covered: the schema documents all parameters at 100%, an output schema exists (so return values need no elaboration), and the description supplies use cases, defaults, and pricing. Minor gaps — no guidance on ordering, no explicit distinction from siblings, no note on the custom-index caveat — but nothing an agent needs to invoke it correctly is missing.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema description coverage is 100%, so the schema carries most parameter semantics, but the description adds decision-relevant information beyond it: the cost-per-1,000-searches for choosing between modes and the default time window. This genuinely helps an agent select mode/value correctly rather than merely restating enum options.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description opens with a specific verb+resource ('Search recent news on any topic across a curated index of news outlets worldwide') and adds the relevance-model criterion. The stated use cases ('current events', 'what outlets reported about a company, person, place or topic') and the return-field list make the tool's distinct job unambiguous versus siblings find_similar and get_contents, even though neither is named.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description clearly states when to use it ('Use it for current events and for what outlets reported about a company, person, place or topic') and gives default behavior (7-day window, fast mode). However, it never names alternatives or states when NOT to use it, such as redirecting to find_similar for similarity matching or get_contents for full article retrieval — so exclusions are left to inference.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
create_chart6 fields changedInput schema / properties / interpretationAdded value: +{
+ "description": "Optional: one of the `alternatives` of a previous call, to chart that reading of the query instead.",
+ "maxLength": 160,
+ "type": "string"
+}Output schema / properties / alternativesAdded value: +{
+ "description": "Other readings of the query: pass one as `interpretation` to chart it instead.",
+ "items": {
+ "type": "string"
+ },
+ "type": "array"
+}Output schema / properties / improvingAdded value: +{
+ "const": true,
+ "description": "Still looking for the exact data: the image and embed update themselves if it arrives.",
+ "type": "boolean"
+}Output schema / properties / rungAdded value: +{
+ "description": "exact (the data asked for), partial (a verified part), related, coverage or stories (the closest thing found).",
+ "type": "string"
+}Output schema / properties / tableAdded value: +{
+ "description": "The rows of a table chart (up to 10), keyed by column label.",
+ "items": {
+ "additionalProperties": {
+ "type": [
+ "string",
+ "number",
+ "null"
+ ]
+ },
+ "propertyNames": {
+ "type": "string"
+ },
+ "type": "object"
+ },
+ "type": "array"
+}Output schema / properties / understoodAdded value: +{
+ "description": "What we understood the user expects to see.",
+ "type": "string"
+}create_chartsearch_news1 field changedInput schema / properties / indexAdded value: +{
+ "description": "Optional: search your own custom index (its id, idx_…) instead of the news index: your own list of URLs, sitemap or feed, created at https://app.typesearch.ai/indexes or with POST /v1/indexes. Same modes and prices; no date filter unless you pass days or dates.",
+ "pattern": "^idx_[a-z0-9]{6,40}$",
+ "type": "string"
+}check_coveragecheck_coveragefind_similarget_contentssearch_newsWorldwide news for agents: search, sync and story clusters across outlets and languages. No key.
Google web and news search for AI agents.
Search global news in natural language. Filter by language, country, date, sentiment, and domain.
Real-time financial news for AI agents: search by ticker and source, with sentiment and entities.
Add one secure layer between your agents and this server.
