imf-mcp-server

Imf Dataframe Query

imf_dataframe_query

Read-onlyIdempotent

Run a read-only SQL SELECT against a DataCanvas table staged by imf_query_dataset. Supports multi-country comparisons, time-series aggregation, and cross-indicator joins. Requires imf_dataframe_describe first to discover table and column names. One SELECT statement per call; a leading WITH … SELECT (CTE) is accepted. DML and DDL are rejected.

Input Schema

Name	Required	Description	Default
`sql`	Yes	Read-only SQL SELECT statement — exactly one statement, starting with SELECT or with a WITH … SELECT common table expression. Reference tables by the names returned by imf_dataframe_describe. Example: SELECT time_period, value FROM spilled_abc123 WHERE time_period >= '2010' ORDER BY time_period.
`canvas_id`	Yes	Canvas ID returned by imf_query_dataset whenever staged=true. Call imf_dataframe_describe with it before writing SQL.

Output Schema

Name	Required	Description
`rows`	No	Largest result-row prefix whose complete structured and formatted response fits the 100,000-character response budget, after the canvas row limit (default 10,000) is applied. Values and column names are always whole here; when a single row only fits with its long cells or column names clipped in the formatted table, listed there one name: value line per column, or replaced there by a one-line pointer to this field, it is returned alone.
`error`	No	Present when the call failed. Absent on success.
`row_count`	No	Number of materialized rows returned in rows. Always equals rows.length and never claims a pre-cap total.
`truncated`	No	True when DataCanvas capped the query at its row limit or the server omitted materialized rows to fit the response-size budget. Page the remainder with a stable ORDER BY plus LIMIT/OFFSET, or narrow the query with WHERE or aggregation.

Schema Changelog

Changes observed during successful MCP inspections.

Changed4 schema fields changedOct 16, 2026

changedOutput schema / properties / error / properties / data / properties / reason / description

Previous value: -"Machine-readable failure mode. Declared by this tool: `canvas_not_found`: canvas_id does not match any registered DataCanvas session (expired, wrong session, or canvas disabled). `missing_table`: The canvas exists but sql references a table that is not staged on it — the table expired, was dropped, or the name is wrong. `invalid_sql`: sql is not a single SELECT statement (a leading WITH … SELECT counts as one), or it is SELECT-shaped but fails to prepare — unknown column, unknown function, or a syntax error. `sql_not_permitted`: sql parses as a SELECT but the read-only gate refuses it — it calls an external-data or PRAGMA table function, reads a system catalog, or plans an operator outside the read-only allowlist. `response_too_large`: The first result row cannot fit in the complete structured and formatted response budget. Other values are possible when a failure originates below the handler."New value: +"Machine-readable failure mode. Declared by this tool: `canvas_not_found`: canvas_id does not match any registered DataCanvas session (expired, wrong session, or canvas disabled). `missing_table`: The canvas exists but sql references a table that is not staged on it — the table expired, was dropped, or the name is wrong. `invalid_sql`: sql is not a single SELECT statement (a leading WITH … SELECT counts as one), or it is SELECT-shaped but fails to prepare — unknown column, unknown function, or a syntax error. `sql_not_permitted`: sql parses as a SELECT but the read-only gate refuses it — it calls an external-data or PRAGMA table function, reads a system catalog, plans an operator outside the read-only allowlist, or names a file instead of a staged table. `sql_execution_error`: sql prepares but fails while it runs, such as a failed cast or an arithmetic overflow. `response_too_large`: The first result row alone exceeds the response budget as structured content, even with the formatted text reduced to a one-line pointer to it. Other values are possible when a failure originates below the handler."

changedOutput schema / properties / error / properties / data / properties / reason / examples

Previous value: -[
-  "canvas_not_found",
-  "missing_table",
-  "invalid_sql",
-  "sql_not_permitted",
-  "response_too_large"
-]New value: +[
+  "canvas_not_found",
+  "missing_table",
+  "invalid_sql",
+  "sql_not_permitted",
+  "sql_execution_error",
+  "response_too_large"
+]

changedOutput schema / properties / rows / description

Previous value: -"Largest result-row prefix whose complete structured and formatted response fits the 100,000-character response budget, after the canvas row limit (default 10,000) is applied."New value: +"Largest result-row prefix whose complete structured and formatted response fits the 100,000-character response budget, after the canvas row limit (default 10,000) is applied. Values and column names are always whole here; when a single row only fits with its long cells or column names clipped in the formatted table, listed there one name: value line per column, or replaced there by a one-line pointer to this field, it is returned alone."

changedOutput schema / properties / rows / items / description

Previous value: -"A result row — keys are the selected column names, values match the column DuckDB types (string, number, null)."New value: +"A result row — keys are the selected column names; values are the JSON form of each DuckDB type: number, string (BIGINT, DECIMAL, and dates included), boolean, null, or an object/array for STRUCT, LIST, and MAP."

Changed2 schema fields changedSep 19, 2026

addedInput schema / properties / canvas_id / pattern
```
Added value: +"^[A-Za-z0-9_-]{10}$"
```

changedOutput schema / properties / error / properties / data / properties / reason / description

Previous value: -"Machine-readable failure mode. Declared by this tool: `canvas_not_found`: canvas_id does not match any registered DataCanvas session (expired, wrong session, or canvas disabled) `missing_table`: The canvas exists but sql references a table that is not staged on it — the table expired, was dropped, or the name is wrong `invalid_sql`: sql is not a single SELECT statement (a leading WITH … SELECT counts as one), or it is SELECT-shaped but fails to prepare — unknown column, unknown function, or a syntax error `sql_not_permitted`: sql parses as a SELECT but the read-only gate refuses it — it calls an external-data or PRAGMA table function, reads a system catalog, or plans an operator outside the read-only allowlist `response_too_large`: The first result row cannot fit in the complete structured and formatted response budget Other values are possible when a failure originates below the handler."New value: +"Machine-readable failure mode. Declared by this tool: `canvas_not_found`: canvas_id does not match any registered DataCanvas session (expired, wrong session, or canvas disabled). `missing_table`: The canvas exists but sql references a table that is not staged on it — the table expired, was dropped, or the name is wrong. `invalid_sql`: sql is not a single SELECT statement (a leading WITH … SELECT counts as one), or it is SELECT-shaped but fails to prepare — unknown column, unknown function, or a syntax error. `sql_not_permitted`: sql parses as a SELECT but the read-only gate refuses it — it calls an external-data or PRAGMA table function, reads a system catalog, or plans an operator outside the read-only allowlist. `response_too_large`: The first result row cannot fit in the complete structured and formatted response budget. Other values are possible when a failure originates below the handler."

Changed5 schema fields changedAug 31, 2026

changedOutput schema / properties / error / properties / data / properties / reason / description

Previous value: -"Machine-readable failure mode. Declared by this tool: `canvas_not_found`: canvas_id does not match any registered DataCanvas session (expired, wrong session, or canvas disabled) `missing_table`: The canvas exists but sql references a table that is not staged on it — the table expired, was dropped, or the name is wrong `invalid_sql`: sql is not a single SELECT statement (a leading WITH … SELECT counts as one), or it is SELECT-shaped but fails to prepare — unknown column, unknown function, or a syntax error `sql_not_permitted`: sql parses as a SELECT but the read-only gate refuses it — it calls an external-data or PRAGMA table function, reads a system catalog, or plans an operator outside the read-only allowlist Other values are possible when a failure originates below the handler."New value: +"Machine-readable failure mode. Declared by this tool: `canvas_not_found`: canvas_id does not match any registered DataCanvas session (expired, wrong session, or canvas disabled) `missing_table`: The canvas exists but sql references a table that is not staged on it — the table expired, was dropped, or the name is wrong `invalid_sql`: sql is not a single SELECT statement (a leading WITH … SELECT counts as one), or it is SELECT-shaped but fails to prepare — unknown column, unknown function, or a syntax error `sql_not_permitted`: sql parses as a SELECT but the read-only gate refuses it — it calls an external-data or PRAGMA table function, reads a system catalog, or plans an operator outside the read-only allowlist `response_too_large`: The first result row cannot fit in the complete structured and formatted response budget Other values are possible when a failure originates below the handler."

changedOutput schema / properties / error / properties / data / properties / reason / examples

Previous value: -[
-  "canvas_not_found",
-  "missing_table",
-  "invalid_sql",
-  "sql_not_permitted"
-]New value: +[
+  "canvas_not_found",
+  "missing_table",
+  "invalid_sql",
+  "sql_not_permitted",
+  "response_too_large"
+]

changedOutput schema / properties / row_count / description

Previous value: -"Number of rows materialized in rows. Equals the canvas row limit when truncated is true — DataCanvas does not report a pre-cap total, so this is never larger than rows.length."New value: +"Number of materialized rows returned in rows. Always equals rows.length and never claims a pre-cap total."

changedOutput schema / properties / rows / description

Previous value: -"Query result rows, capped at the canvas row limit (default 10,000)."New value: +"Largest result-row prefix whose complete structured and formatted response fits the 100,000-character response budget, after the canvas row limit (default 10,000) is applied."

changedOutput schema / properties / truncated / description

Previous value: -"True when the query matched more rows than the canvas row limit and the result was capped. Page the remainder with a stable ORDER BY plus LIMIT/OFFSET, or narrow the query with WHERE or aggregation."New value: +"True when DataCanvas capped the query at its row limit or the server omitted materialized rows to fit the response-size budget. Page the remainder with a stable ORDER BY plus LIMIT/OFFSET, or narrow the query with WHERE or aggregation."

Changed1 schema field changedAug 31, 2026

changedInput schema / properties / canvas_id / description

Previous value: -"Canvas ID returned by imf_query_dataset when results were too large for inline delivery."New value: +"Canvas ID returned by imf_query_dataset whenever staged=true. Call imf_dataframe_describe with it before writing SQL."

Changed6 schema fields changedAug 25, 2026

changedInput schema / $schema

Previous value: -"http://json-schema.org/draft-07/schema#"New value: +"https://json-schema.org/draft/2020-12/schema"

addedInput schema / additionalProperties
```
Added value: +false
```

changedOutput schema / $schema

Previous value: -"http://json-schema.org/draft-07/schema#"New value: +"https://json-schema.org/draft/2020-12/schema"

addedOutput schema / anyOf

Added value: +[
+  {
+    "not": {
+      "required": [
+        "error"
+      ]
+    },
+    "required": [
+      "rows",
+      "row_count",
+      "truncated"
+    ]
+  },
+  {
+    "required": [
+      "error"
+    ]
+  }
+]

addedOutput schema / properties / error

Added value: +{
+  "additionalProperties": {},
+  "description": "Present when the call failed. Absent on success.",
+  "properties": {
+    "code": {
+      "description": "JSON-RPC error code for this failure.",
+      "maximum": 9007199254740991,
+      "minimum": -9007199254740991,
+      "type": "integer"
+    },
+    "data": {
+      "additionalProperties": {},
+      "properties": {
+        "reason": {
+          "description": "Machine-readable failure mode. Declared by this tool: `canvas_not_found`: canvas_id does not match any registered DataCanvas session (expired, wrong session, or canvas disabled) `missing_table`: The canvas exists but sql references a table that is not staged on it — the table expired, was dropped, or the name is wrong `invalid_sql`: sql is not a single SELECT statement (a leading WITH … SELECT counts as one), or it is SELECT-shaped but fails to prepare — unknown column, unknown function, or a syntax error `sql_not_permitted`: sql parses as a SELECT but the read-only gate refuses it — it calls an external-data or PRAGMA table function, reads a system catalog, or plans an operator outside the read-only allowlist Other values are possible when a failure originates below the handler.",
+          "examples": [
+            "canvas_not_found",
+            "missing_table",
+            "invalid_sql",
+            "sql_not_permitted"
+          ],
+          "type": "string"
+        },
+        "recovery": {
+          "additionalProperties": {},
+          "description": "Actionable next step for the caller.",
+          "properties": {
+            "hint": {
+              "type": "string"
+            }
+          },
+          "required": [
+            "hint"
+          ],
+          "type": "object"
+        },
+        "retryable": {
+          "description": "Whether retrying may succeed.",
+          "type": "boolean"
+        }
+      },
+      "type": "object"
+    },
+    "message": {
+      "description": "Human-readable description of what went wrong.",
+      "type": "string"
+    }
+  },
+  "required": [
+    "code",
+    "message"
+  ],
+  "type": "object"
+}

removedOutput schema / required

Removed value: -[
-  "rows",
-  "row_count",
-  "truncated"
-]

Changed4 schema fields changedAug 6, 2026

changedInput schema / properties / sql / description

Previous value: -"Read-only SQL SELECT statement. Must start with SELECT. Reference tables by the names returned by imf_dataframe_describe. Example: SELECT time_period, value FROM spilled_abc123 WHERE time_period >= '2010' ORDER BY time_period."New value: +"Read-only SQL SELECT statement — exactly one statement, starting with SELECT or with a WITH … SELECT common table expression. Reference tables by the names returned by imf_dataframe_describe. Example: SELECT time_period, value FROM spilled_abc123 WHERE time_period >= '2010' ORDER BY time_period."

changedOutput schema / properties / row_count / description

Previous value: -"Total matching rows before the cap — may exceed rows.length."New value: +"Number of rows materialized in rows. Equals the canvas row limit when truncated is true — DataCanvas does not report a pre-cap total, so this is never larger than rows.length."

addedOutput schema / properties / truncated

Added value: +{
+  "description": "True when the query matched more rows than the canvas row limit and the result was capped. Page the remainder with a stable ORDER BY plus LIMIT/OFFSET, or narrow the query with WHERE or aggregation.",
+  "type": "boolean"
+}

changedOutput schema / required

Previous value: -[
-  "rows",
-  "row_count"
-]New value: +[
+  "rows",
+  "row_count",
+  "truncated"
+]

First observedJun 7, 2026

TDQS

A4.2/5.0

Behavior4/5

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

Annotations already declare readOnlyHint/idempotentHint, but the description adds enforced behavior beyond them: DML and DDL are rejected, one SELECT statement per call, and leading WITH…SELECT CTEs are accepted. These are concrete operational constraints an agent needs. It stops short of stating row/size limits or error behavior.

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

Conciseness5/5

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

Four tight sentences, each carrying distinct information (operation, capabilities, prerequisite, statement constraints), with the core action front-loaded. No filler or restatement of the title.

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

Completeness4/5

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

Covers the staging→describe→query workflow and the SQL dialect constraints, and an output schema exists so return values need not be explained. It omits result-size limits or cost/rate considerations that would fully round out a query tool.

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 both parameters are already fully documented in the schema, including the canvas_id pattern and the sql example. The description adds no syntax or format detail beyond the schema, so the baseline 3 applies.

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?

States a specific verb+resource ('Run a read-only SQL SELECT against a DataCanvas table') and names the sibling that creates the table it operates on. The scope statement ('staged by imf_query_dataset') lets an agent distinguish it from imf_query_dataset and imf_dataframe_describe 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.

Usage Guidelines4/5

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

Explicit prerequisite ('Requires imf_dataframe_describe first to discover table and column names') and supported use cases (multi-country comparisons, time-series aggregation, cross-indicator joins). It does not name ineligible alternatives such as imf_get_database, so it stops just short of full when/when-not routing.

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

All imf-mcp-server tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Imf Dataframe Query

Input Schema

Output Schema

Schema Changelog

TDQS

Glama MCP Gateway

Resources