Argentine economic statistics sourced from official data (BCRA, INDEC, datos.gob.ar) as tools for AI agents: exchange rates, inflation, rates, reserves, activity, wages and trade. Search series, fetch history, projections with calibrated bands and the BCRA market consensus, weekly summary and release calendar. Independent project, not affiliated with any government agency. Free, attribution required.
Scored across 5 tools
Each tool has a clearly distinct purpose: search_series is the entry point for discovery, get_series returns past/current values, get_projection returns future projections, get_release_calendar lists scheduled releases, and get_weekly_summary provides a weekly overview. The descriptions explicitly cross-reference each other (e.g., 'use get_series for past values'), eliminating overlap. No two tools appear to do the same thing.
All tool names follow a consistent snake_case verb_noun pattern: get_projection, get_release_calendar, get_series, get_weekly_summary, search_series. The only variation is 'search_' vs 'get_', which is semantically appropriate and still uses the same naming convention. There is no mixing of camelCase or inconsistent verb styles.
Five tools is well-scoped for a read-only economic data API: one for discovery, one for history, one for projections, one for release dates, and one for a weekly digest. Each tool has a clear reason to exist and there is no redundancy. The count is neither too thin nor too heavy.
The surface covers the core lifecycle—find a series, get its history, project it forward, check upcoming releases, and read a weekly summary. Minor gaps exist, such as no bulk retrieval of multiple series at once and no browse-all-categories endpoint beyond keyword search. These are workable around, but they slightly limit completeness.
Project a series forward: point values with 80% and 95% bands, the models combined (ETS, ARIMA, Theta, chosen by walk-forward backtest), backtest errors and, when the series has one, the BCRA market-expectations survey (REM) median for the same months. Use it for "what's next" questions about inflation, exchange rates, interest rates or reserves; use get_series for past values. It is a statistical reference, not a forecast or advice: say so when you quote it. Needs a key from search_series. Series with too little history return an error. Read-only, no side effects, no authentication. Rate limited to 60 calls per minute per IP. Data updates at most once per business day.
| Name | Required | Description | Default |
|---|---|---|---|
| key | Yes | Series key as returned by search_series, e.g. "bcra:27". | |
| horizon | No | How far ahead: calendar days for daily series (default 30, up to 183), periods for the rest (default 12 months, 8 quarters). |
| Name | Required | Description |
|---|---|---|
| key | Yes | |
| name | Yes | |
| unit | Yes | |
| bands | Yes | |
| engine | Yes | lote: daily batch models; sitio: computed on request |
| method | Yes | |
| horizon | Yes | |
| members | No | |
| backtest | Yes | |
| benchmark | Yes | |
| consensus | No | |
| disclaimer | Yes | |
| attribution | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations cover the safety profile, but the description adds genuinely new operational context: rate limit of 60 calls/minute per IP, no authentication required, data refreshed at most once per business day, and a documented error condition for short series. It also sets a communication obligation ('not a forecast or advice: say so when you quote it'), which no structured field conveys.
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?
Front-loaded with the capability and outputs before usage, prerequisites and caveats. Dense and mostly waste-free, though some clauses ('Read-only, no side effects, no authentication') restate what annotations already declare.
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?
Despite an output schema being present, the description usefully characterizes the returned content, and it covers the prerequisites, error behavior and freshness constraints an agent needs before calling. Nothing material 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% and both parameters are documented in the schema, including the horizon default/units and the example key format, so the baseline is 3. The description only reinforces key provenance ('Needs a key from search_series') without adding new syntax or unit detail for horizon.
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?
States a specific verb and resource ('Project a series forward') and immediately enumerates the output the tool produces (point values, 80%/95% bands, combined ETS/ARIMA/Theta models, backtest errors, REM median). It explicitly distinguishes itself from the sibling get_series, so an agent can route without opening either schema.
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?
Gives explicit when-to-use ('what's next' questions about inflation, exchange rates, interest rates or reserves) and names the alternative for the opposite case ('use get_series for past values'). It also states the prerequisite ('Needs a key from search_series') and the failure condition (too little history returns an error).
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List the scheduled INDEC releases (CPI inflation, EMAE activity, wages, trade, industry, construction, employment, poverty) between two dates, each with date, time in Argentina, report name, period covered and the series keys it updates. Use it to answer "when is the next inflation figure out" or to know whether a value is about to change. Defaults to the next 60 days from today. Only INDEC for now. Read-only, no side effects, no authentication. Rate limited to 60 calls per minute per IP. Data updates at most once per business day.
| Name | Required | Description | Default |
|---|---|---|---|
| to | No | Last day to include, YYYY-MM-DD. Defaults to 60 days after from. | |
| from | No | First day to include, YYYY-MM-DD, Argentina time. Defaults to today. |
| Name | Required | Description |
|---|---|---|
| to | Yes | |
| url | Yes | |
| from | Yes | |
| releases | Yes | |
| time_zone | Yes | |
| attribution | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations cover read-only/idempotent/non-destructive, but the description adds contextual traits annotations cannot express: no authentication required, a 60 calls/minute per-IP rate limit, and a data-refresh cadence of at most once per business day. These are exactly the operational facts an agent needs before scheduling repeated calls.
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?
A single dense paragraph that front-loads the payload and the return fields before falling back to defaults, limitations, and operational notes. Every sentence carries information, though the operational tail could be marginally tightened.
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 explanation isn't required, yet the description still previews the returned fields. Combined with coverage of defaults, source limitation, auth, rate limits, and refresh cadence, nothing an agent needs to call and interpret this tool 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 both parameters (from/to patterns and defaults) are fully documented in the schema itself. The description only restates the 'between two dates' semantics and the 60-day default, adding no format or edge-case detail beyond the schema. Baseline 3 is appropriate.
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?
States a specific verb (List) and resource (scheduled INDEC releases) with explicit scope (between two dates) and enumerates the report families and returned fields. An agent can tell this apart from get_series/search_series, which return data rather than a forward-looking schedule.
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?
Gives concrete usage scenarios ('when is the next inflation figure out', 'know whether a value is about to change') and states the default window and the 'only INDEC for now' limitation. It stops short of naming sibling alternatives (e.g., get_series for values) or stating when not to use it.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return the historical values of one series with its metadata (name, unit, frequency, source, page URL, last date and value). Use it for past and current values; use get_projection for future ones. Needs a key from search_series; well-known keys: bcra:27 monthly inflation, bcra:28 annual inflation, bcra:4 retail USD rate, bcra:5 wholesale USD rate, bcra:1 reserves, bcra:7 BADLAR rate, bcra:44 TAMAR rate, bcra:31 UVA, bcra:30 CER. Without range or from it returns the last year for daily series and the last five years for the rest. At most 500 points, the most recent ones; a note says how many older points were left out. An unknown key returns an error that asks to search first. Read-only, no side effects, no authentication. Rate limited to 60 calls per minute per IP. Data updates at most once per business day.
| Name | Required | Description | Default |
|---|---|---|---|
| key | Yes | Series key with its source prefix, as returned by search_series: "bcra:27", "datos:143.3_NO_PR_2004_A_31", "derivada:brecha_blue". | |
| from | No | Start date as YYYY-MM or YYYY-MM-DD; rounded down to the first day of that month. Takes precedence over range. | |
| range | No | Time window ending at the last available value: 1m, 3m, 6m, 1y, 5y or max (full history). Ignored if from is given. |
| Name | Required | Description |
|---|---|---|
| key | Yes | |
| url | Yes | |
| name | Yes | |
| note | No | |
| unit | Yes | |
| points | Yes | [date YYYY-MM-DD, value], oldest first |
| source | Yes | |
| category | Yes | |
| frequency | Yes | D daily, M monthly, T quarterly, S semiannual, A annual |
| last_date | Yes | |
| last_value | Yes | |
| attribution | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already establish read-only/idempotent/non-destructive, and the description goes well beyond them: default windows (1 year daily, 5 years otherwise), the 500-point cap with a truncation note, error behavior for unknown keys, 60 calls/min per IP rate limit, and once-per-business-day refresh cadence.
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?
Front-loaded with purpose, then routing, then defaults and limits - a logical order with no filler. It is dense but each clause carries distinct information (defaults, cap, error, rate limit), though the single long block is slightly heavy for a 3-parameter tool.
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?
Covers everything an agent needs before calling: required key provenance, default windows, result cap and truncation signaling, error semantics, rate limits, and auth requirements. With an output schema present, it correctly refrains from enumerating return fields in detail.
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 coverage is 100%, so the baseline is 3, but the description adds value the schema lacks: the default behavior when neither from nor range is supplied, and concrete well-known key examples beyond the schema's sample values. It does not restate the from/range precedence rule that the schema already carries.
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?
States a specific verb (return) and resource (historical values of one series) plus the metadata set returned. It explicitly distinguishes itself from get_projection ("use get_projection for future ones"), so an agent can route correctly without reading either schema.
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?
Gives explicit when-to-use (past/current values), when-to-use-the-alternative (future values -> get_projection), and a prerequisite (a key from search_series). It even supplies well-known keys, which is unusually actionable guidance.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return one plain-Spanish sentence per indicator (USD retail and wholesale, reserves, interest rates, inflation, UVA) describing how it moved during a closed week, with direction and links, plus the permalink of that week's summary page. Use it for a quick "how did the Argentine economy do this week" overview; use get_series for exact values. Defaults to the last closed week (Monday to Sunday). Read-only, no side effects, no authentication. Rate limited to 60 calls per minute per IP. Data updates at most once per business day.
| Name | Required | Description | Default |
|---|---|---|---|
| week_ending | No | Sunday that closes the week, as YYYY-MM-DD. Omit for the last closed week; any other date is anchored to the previous Sunday. |
| Name | Required | Description |
|---|---|---|
| url | Yes | |
| week | Yes | |
| items | Yes | |
| attribution | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations cover safety (readOnly, idempotent, non-destructive, closed-world), and the description goes well beyond them with a rate limit (60/min per IP), a data-freshness cadence (at most once per business day), and the default time anchor. These are operationally relevant traits an agent cannot get from 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.
Is the description appropriately sized, front-loaded, and free of redundancy?
Output shape is front-loaded, followed by usage, default, safety, limits, and freshness in that order. Every sentence carries distinct information; the only near-redundancy is restating read-only/no-side-effects already in annotations, which is minor.
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 single-optional-parameter read tool with an output schema, annotations, and full schema coverage, the definition supplies everything needed: scope, alternative, default window, rate limit, and refresh cadence. Nothing an agent needs to call 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 coverage is 100% and the schema already documents week_ending's pattern, the omit-default, and Sunday anchoring, so the baseline is 3. The description adds the clarifying fact that the default week runs Monday to Sunday, which sharpens what 'closed week' means beyond the schema wording.
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?
States a precise verb (return) and resource (one plain-Spanish sentence per indicator for a closed week, plus the week's permalink) and enumerates the indicators covered. It also explicitly distinguishes itself from the sibling get_series, so an agent can route without opening schemas.
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?
Gives an explicit use case ('quick how did the Argentine economy do this week overview') and names the alternative with the condition that selects it ('use get_series for exact values'). No inference required.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Find Argentine economic time series by keywords and get their keys. Call this first whenever you don't already know a series key, then pass the key to get_series (history) or get_projection (forecast). Covers BCRA (exchange rates, reserves, interest rates, monetary aggregates, UVA, CER), INDEC via datos.gob.ar (inflation, activity, wages, trade, employment, poverty) and derived series (real rate, FX gap). Matches every word, ignoring accents and case; series names are in Spanish, so Spanish keywords work best ("inflacion", "dolar", "reservas", "badlar", "salarios"). Returns up to 15 matches with key, name, category, source and page URL; an empty list means no match, not an error. Read-only, no side effects, no authentication. Rate limited to 60 calls per minute per IP. Data updates at most once per business day.
| Name | Required | Description | Default |
|---|---|---|---|
| query | Yes | Keywords to match against series names and categories, e.g. "inflacion mensual" or "tipo de cambio". |
| Name | Required | Description |
|---|---|---|
| note | No | |
| query | Yes | |
| results | Yes | |
| attribution | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnly/idempotent/non-destructive, and the description adds substantial context beyond them: word-based matching that ignores accents and case, a 60 calls/minute per IP rate limit, no authentication, and an at-most-daily data refresh. These are exactly the operational traits an agent needs and none are derivable from 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.
Is the description appropriately sized, front-loaded, and free of redundancy?
Front-loads purpose and the call-this-first directive, then layers coverage, matching behavior, return shape and limits. Every sentence earns its place; nothing is padding despite the density.
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 single-parameter discovery tool with an output schema, the description is complete: it explains coverage scope, matching rules, result cardinality and fields, error semantics, rate limits and freshness. An agent has everything needed to invoke and interpret 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 coverage is 100%, so the query parameter is already documented, but the description adds real meaning: matching is every-word, accent- and case-insensitive, series names are Spanish so Spanish keywords work best, with concrete examples ('inflacion', 'dolar', 'reservas'). This goes beyond the schema's single example and helps the agent formulate a better query.
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?
States a specific verb and resource ('Find Argentine economic time series by keywords and get their keys') and explicitly positions itself relative to siblings get_series and get_projection. An agent can distinguish this discovery tool from the history/forecast tools without opening a schema.
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?
Gives explicit when-to-use guidance ('Call this first whenever you don't already know a series key') and names the downstream alternatives with their purpose. It also covers the negative case: an empty list means no match, not an error, so the agent knows not to retry or treat it as failure.
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.
buscar_seriescalendarioget_projectionget_release_calendarget_seriesget_weekly_summaryobtener_serieproyeccionresumen_semanalsearch_seriesbuscar_seriescalendarioobtener_serieproyeccionresumen_semanal
Add one secure layer between your agents and this server.