Find air-quality monitoring stations, read latest sensor values, and pull historical pollutant series via MCP. STDIO or Streamable HTTP.
Strong tool definitions with excellent descriptions (150-350 chars, well above baseline 194 avg) and comprehensive parameter documentation. All 6 tools have clear verb-noun naming (openaq_*) and detailed input schemas with types and descriptions. Output schemas are documented in descriptions. Key strength: parameter descriptions include format guidance, constraints, and dependency hints (e.g., 'Requires DataCanvas to be enabled', 'Pair with radius for a near-me search'). Minor gaps: no explicit error handling guidance in descriptions, no per-tool permission declarations, and output field structures not formally documented in JSON Schema format. Tool composition is excellent, each tool has one clear responsibility and outputs chain naturally (find_locations → get_readings/get_measurements → dataframe_query).
List the tables and columns staged on a DataCanvas so you can write valid SQL for openaq_dataframe_query without guessing column names. Returns each measurement table (measurements_<sensorId>) with its row count and column names. Requires DataCanvas to be enabled.
Run a read-only SQL SELECT against the measurement tables openaq_get_measurements staged on a DataCanvas. Reference tables by the name the measurements call returned (measurements_<sensorId>). For aggregation (monthly means, exceedance counts) and cross-sensor comparison over series too large to inline. Only SELECT is allowed — writes, DDL, and file/network table functions are rejected. Responses carry at most 200 rows; aggregate in SQL, or page with ORDER BY plus LIMIT/OFFSET, rather than selecting a whole table.
Find air-quality monitoring stations (measured by physical sensors, not modeled) near a point, within a bounding box, or by country. Returns each station's id, name, coordinates, distance from the query point (when searching by coordinates), country, provider, the parameters its sensors measure, and the timestamp of its most recent data (datetimeLast). Required first step: openaq_get_readings and openaq_get_measurements key on the location id this returns. Coverage is uneven and real — a station only reports the parameters it measures, and the absence of a nearby station means no monitoring there, not clean air. For dense modeled coverage anywhere on Earth, use open-meteo-mcp-server's air-quality tool instead.
Output schemas not formally documented in JSON Schema format. Descriptions explain what fields are returned (e.g., 'Returns each station's id, name, coordinates...') but no structured schema definition visible in code.
No error handling guidance in tool descriptions. Descriptions do not explain what errors can occur, how to recover, or what the LLM should do if a call fails (e.g., 'If no stations found within radius, try increasing radius or using bbox').
No permission declarations. Tools do not state required scopes (e.g., 'read:openaq-api'). Limits audit trail clarity and least-privilege configuration.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 69 | 2026-07-28+ | v2 |
Historical measurement series for one pollutant at one station over a date range — for trend analysis and "was last week worse than the monthly average?". Pass a locationId and a parametersId and work in stations — you get the series for that pollutant at that station. Choose aggregation: raw (every reported value), hourly, or daily — daily and hourly add a per-bucket statistical summary (min, median, max, mean, sd). Large ranges produce thousands of rows and stage on a DataCanvas: the response returns a preview plus a canvasId and table name — call openaq_dataframe_describe on the canvasId for the table's columns, then openaq_dataframe_query to run SQL over it. Passing a canvas_id stages the series there whatever its size, so two stations land on one canvas for a side-by-side comparison. Values carry their unit; the server never converts between µg/m³, ppm, and ppb.
Latest measured value for every sensor at a monitoring station — the current-conditions tool. Returns one record per parameter, each with the value, its unit, the UTC and local timestamp, and the sensor id, joined so every value carries its pollutant and unit (the raw latest feed is keyed only by sensor id). Pass a locationId from openaq_find_locations, or pass coordinates to auto-resolve to the nearest station that measures the requested parametersId. Data recency varies by station reporting cadence — read each value's timestamp to know whether "latest" is minutes or hours old. These are measured observations with coverage gaps, not a modeled grid.
Catalog of country-level coverage: id, ISO code, name, the date span of available station data (datetimeFirst/datetimeLast), and which parameters are measured anywhere in that country. The availability check before a regional sweep — answers "which countries have NO2 monitoring?" and tells you whether a country has recent data before you call openaq_find_locations. Coverage is uneven worldwide; this surfaces where measured data exists.
Parameter 'limit' in openaq_find_locations and openaq_list_countries lacks explicit min/max bounds in schema. Description states '1 - 100' and '1 - 20' but JSON Schema should enforce via minimum/maximum properties.
openaq_get_measurements 'aggregation' enum is well-defined, but no guidance on when to use 'raw' vs 'hourly' vs 'daily' beyond 'Use daily for multi-month trends'. Could add: 'raw for fine-grained recent analysis (last 7 days), hourly for weekly trends, daily for monthly+ trends.'